AI Skills for Business Competency Framework
Abstract
Artificial intelligence (AI) is reshaping how people live, work, and participate in society, offering significant potential to enhance productivity, competitiveness, and public value. As AI systems become increasingly embedded in everyday services and organisational processes, employees across all roles require the skills and confidence to engage with these technologies safely, effectively, and ethically. For organisations, successful AI adoption depends not only on technical capability but also on ensuring alignment with ethical principles, regulatory expectations, and long-term public trust. The AI Skills for Business Competency Framework has been developed to provide a structured understanding of the skills needed across the workforce. It offers a clear, role-aligned articulation of the knowledge, skills, and behaviours required for responsible and effective AI engagement across the workplace and wider society. The framework is designed to support organisations, educators, and policymakers in planning capability development, informing curriculum design, guiding organisational governance, and targeting professional training. The AI Skills for Business Competency Framework was formally launched in May 2024 in response to the UK Government National AI Strategy commitment to "publish research into what skills are needed to enable employees to use AI in a business setting and identify how national skills provision can meet these needs. Since its launch, the framework has been used to underpin national skills infrastructures, support supply-demand-gap analysis for AI skills in the UK, and support the development of national curricula to meet the needs of industry. We now bring to the community a consultation draft for Version 3 of the framework, the culmination of 12 months of engagement and co-design with businesses, educators, and policymakers. The new edition of the framework presents substantial updates across competency domains, new duty-based articulations of AI skills. We recognise that no two organisations are alike, and that different sectors have very different needs. That’s why this consultation process is so important. It is your opportunity to tell us what resonates, to challenge us on what needs to change or be improved, and to tell us what “good” looks like for your sector, your staff, and your customers. Please provide your feedback via the Consultation Feedback Form.
Full text
1 AI Skills for Business Competency Framework Version 3 (Consultation Draft) 8th December 2025
Consultation Draft for Review (December 2025). 2 Introduction Artificial intelligence (AI) is reshaping how people live, work, and participate in society, offering significant potential to enhance productivity, competitiveness, and public value. As AI systems become increasingly embedded in everyday services and organisational processes, employees across all roles require the skills and confidence to engage with these technologies safely, effectively, and ethically. For organisations, successful AI adoption depends not only on technical capability but also on ensuring alignment with ethical principles, regulatory expectations, and long-term public trust. The AI Skills for Business Competency Framework has been developed to provide a structured understanding of the skills needed across the workforce. It offers a clear, rolealigned articulation of the knowledge, skills, and behaviours required for responsible and effective AI engagement across the workplace and wider society. The framework is designed to support organisations, educators, and policymakers in planning capability development, informing curriculum design, guiding organisational governance, and targeting professional training. The AI Skills for Business Competency Framework was formally launched in May 2024 in response to the UK Government National AI Strategy commitment to "publish research into what skills are needed to enable employees to use AI in a business setting and identify how national skills provision can meet these needs. Since its launch, the framework has been used to underpin national skills infrastructures, support supply-demand-gap analysis for AI skills in the UK, and support the development of national curricula to meet the needs of industry. We now bring to the community a consultation draft for Version 3 of the framework, the culmination of 12 months of engagement and co-design with businesses, educators, and policymakers. The new edition of the framework presents substantial updates across competency domains, new duty-based articulations of AI skills. We recognise that no two organisations are alike, and that different sectors have very different needs. That’s why this consultation process is so important. It is your opportunity to tell us what resonates, to challenge us on what needs to change or be improved, and to tell us what “good” looks like for your sector, your staff, and your customers. For details on how to contribute to the consultation, please visit https://www.turing.ac.uk/skills/collaborate/ai-skills-business-framework/ and contact [email protected] if you have any questions.
Consultation Draft for Review (December 2025). 3 Table of Contents Introduction .............................................................................................................................. 2 Table of Contents ..................................................................................................................... 3 Structure of the Framework ..................................................................................................... 8 Personas .............................................................................................................................. 8 Duties ................................................................................................................................... 8 Competencies ...................................................................................................................... 8 Proficiency Levels ................................................................................................................ 9 AI Project Lifecycle Stages ................................................................................................. 10 Presentation of the Final Framework ................................................................................. 10 Personas ................................................................................................................................ 11 Citizen ................................................................................................................................ 11 Worker ................................................................................................................................ 11 Professional ........................................................................................................................ 11 Leader ................................................................................................................................ 11 Citizen Advocate ................................................................................................................ 12 Duties ..................................................................................................................................... 13 Citizen ................................................................................................................................ 13 Worker ................................................................................................................................ 14 Professional ........................................................................................................................ 16 Leader ................................................................................................................................ 19 Citizen Advocate ................................................................................................................ 20 Summary of Competencies .................................................................................................... 22 Competencies and Knowledge, Skills and Behaviours (KSBs) Statements .......................... 26 C1: Identifies where AI operates in daily life and understands its basic functions and limits. ........................................................................................................................................... 26 C2: Engages with AI-driven tools and services thoughtfully, understanding both benefits and risks. ................................................................................................................................... 26 C3: Understands user rights, choices and complaint routes when AI affects access, pricing, or critical services and opportunities. ................................................................................. 27 C4: Understands how personal data is collected, processed, and used to train and personalise AI systems. ..................................................................................................... 27 C5: Takes active steps to protect personal data and manage privacy across AI-enabled platforms. ............................................................................................................................ 27 C6: Shares only what is necessary, understands consent choices, and reviews permissions regularly. ............................................................................................................................. 28 C7: Demonstrates awareness of AI-generated material and understands its implications for information integrity. ........................................................................................................... 28 C8: Critically assesses the reliability and fairness of AI-generated outputs before accepting or sharing them. ................................................................................................................. 28 C9: Demonstrates awareness of the risks of sharing AI-generated content by pausing,
Consultation Draft for Review (December 2025). 4 checking context, and adding sources when appropriate. ................................................. 29 C10: Detects and avoids harmful or deceptive uses of AI, including fraud, manipulation, or misinformation. ................................................................................................................... 29 C11: Recognises and responds to instances of unfairness, exclusion, or bias in AI systems and promotes equitable practices. ..................................................................................... 30 C12: Knows where and how to report AI-related scams, abuse, or unfair outcomes and helps others do the same. .................................................................................................. 30 C13: Contributes to conversations about AI with respect and evidence. ........................... 30 C14: Contributes to AI related discussions by actively listening, sharing perspectives and respecting diverse experience. ........................................................................................... 31 C15: Makes space for different experiences and needs so public conversations about AI are fair and inclusive. ......................................................................................................... 31 C16: Comply with data protection policies and safeguard sensitive information when using AI systems. ......................................................................................................................... 32 C17: Use AI-powered platforms confidently and effectively in everyday work. .................. 32 C18: Explore and experiment with AI tools responsibly, ensuring alignment with organisational policies. ....................................................................................................... 33 C19: Demonstrate and promote ethical awareness in everyday AI use by respecting privacy, questioning outputs, and encouraging responsible practices among colleagues. ............. 33 C20: Use AI safely and inclusively in collaborative settings, promoting transparency and equitable participation, with awareness of potential bias and accessibility limitations. ...... 34 C21: Be transparent about the role of AI in producing work outputs. ................................. 34 C22: Apply AI tooling into workflows responsibly, ensuring alignment with organisational policies and practices. ........................................................................................................ 35 C23: Interpret AI-driven outputs critically and apply professional judgement. .................... 35 C24: Adapt to the changing nature of work through ongoing engagement with AI tools. .. 36 C25: Contribute to organisational learning about AI use. ................................................... 36 C26: Serve as a local advocate for the effective and ethical use of AI tools and practices within their team or department. ......................................................................................... 37 C27: Provide first-line guidance and support to co-workers on appropriate AI use and troubleshooting common issues. ........................................................................................ 37 C28: Encourage compliance with organisational AI policies and data governance standards, escalating risks or concerns to management when necessary. ......................................... 38 C29: Identify areas where AI can support efficiency, innovation, or service improvements, and share recommendations with leadership. .................................................................... 38 C30: Encourage participation in AI training, sharing resources and best practice examples with colleagues. .................................................................................................................. 39 C31: Act as a liaison between frontline staff and senior leadership, feeding back on staff needs, adoption challenges, and success stories. ............................................................. 39 C32: Keep up to date with developments in AI tools and organisational policy, ensuring colleagues are aware of relevant updates and guidance. .................................................. 40 C33: Demonstrate effective use of AI by modelling good practice of AI use in day-to-day tasks, showcasing practical and responsible applications that inspire peers. .................... 40 C34: Demonstrate ownership of personal development and engage in continuous learning activities. ............................................................................................................................. 40
Consultation Draft for Review (December 2025). 5 C35: Design and develop data-driven solutions and AI systems. ...................................... 41 C36: Test and monitor AI systems to assess performance and accuracy. ........................ 42 C37: Ensure that AI systems are designed and developed in an ethical, safe and responsible way. .................................................................................................................................... 43 C38: Project and change management through delivery of innovative projects. ............... 44 C39: Apply rigorous methodology to generate, validate and share knowledge that supports responsible AI practices. .................................................................................................... 44 C40: Design and maintains data architectures, pipelines, and infrastructure that support reliable, secure, and scalable AI operations. ...................................................................... 45 C41: Evaluate human-AI collaboration and ensure AI systems are aligned with user needs. ........................................................................................................................................... 46 C42: Contribute to the design, development and deployment of AI systems in alignment with defined purpose, policy and ethical standards that include safety, robustness and reliability. ........................................................................................................................................... 46 C43: Uphold professional standards in data stewardship, model evaluation and stakeholder engagement. ...................................................................................................................... 47 C44: Design AI architectures that align technical, ethical, and organisational objectives. . 48 C45: Effectively select, integrate, and contribute to open-source tools, frameworks, and libraries that support robust and reproducible AI system development. ............................. 48 C46: Engage with open-source communities in ways that uphold ethical, secure, and inclusive practices, promoting responsible participation in the AI ecosystem. ................... 49 C47: Collect and prepare the required datasets in line with organisation standards, ensuring accuracy, compliance and readiness for AI project use. .................................................... 50 C48: Support the application of AI and related technologies in ways that uphold rights, promote trust and deliver value. ......................................................................................... 51 C49: Collaborate effectively in multidisciplinary AI development teams. ........................... 52 C50: Implement, customise, and optimise low-code and no-code solutions that address organisational challenges, improve workflows, and enhance efficiency. ........................... 52 C51: Advocate for and ensure automation solutions are implemented responsibly, with fairness, transparency, and consideration for workforce wellbeing. ................................... 53 C52: Develop and apply methods to ensure AI explainability and interpretability to ensure that AI systems are transparent, interpretable, and understandable to appropriate audiences. .......................................................................................................................... 54 C53: Design and operate AI systems with due consideration for environmental and social sustainability. ...................................................................................................................... 54 C54: Apply AI governance and risk management principles to ensure compliance and accountability. ..................................................................................................................... 55 C55: Support the development of AI systems to proactively manage potential workforce impacts. .............................................................................................................................. 56 C56: Support the responsible, ethical and safe adoption of AI. ......................................... 56 C57: Ensure responsible governance, security and stewardship of data across the AI lifecycle. .............................................................................................................................. 57 C58: Support documentation and auditability of AI systems and tooling. .......................... 58 C59: Deploy data-driven and AI solutions and integrate them with organisation's systems. ........................................................................................................................................... 58
Consultation Draft for Review (December 2025). 6 C60: Maintain and continuously improve deployed AI systems. ........................................ 60 C61: Ensure the security, robustness, and resilience of AI systems. ................................. 60 C62: Contribute to the ongoing evaluation and refinement of solutions to ensure they remain effective and responsible, and aligned with evolving contexts. .......................................... 61 C63: Performance management and continuous improvement. ........................................ 62 C64: Engage senior leadership, end-users and relevant stakeholders across all phases of the Data and AI Lifecycle to ensure systems and data-driven solutions are fit for purpose. ........................................................................................................................................... 62 C65: Support leadership and advocate for responsible introduction of artificial intelligence solutions. ............................................................................................................................ 63 C66: Communicate effectively with a range of audiences, ensuring clarity, transparency, and timeliness in reporting progress on AI throughout the AI project lifecycle. .................. 63 C67: Design and develop data-driven and AI solutions that are intuitive and human-centred. ........................................................................................................................................... 64 C68: Collaborate across disciplines to identify risks, integrate diverse perspectives and support responsible innovation. .......................................................................................... 65 C69: Research, critically evaluate, and apply emerging knowledge to AI automation developments, ensuring practice remains current, ethical, and forward-looking. ............... 66 C70: Evaluate emerging AI technologies and their potential societal impact. .................... 67 C71: Support learning, mentorship, and capacity building in responsible and ethical AI practice. .............................................................................................................................. 67 C72: Engage with and/or develop governance processes that enable validation, oversight and escalation. ................................................................................................................... 68 C73: Evaluate and oversee external AI vendors and third-party tools to ensure technical fitness, compliance, and responsible operation. ................................................................ 69 C74: Define and communicate a coherent AI strategy that aligns with organisational goals and values. ......................................................................................................................... 69 C75: Balance innovation with practical delivery in AI adoption. ......................................... 70 C76: Sponsor governance frameworks that safeguard lawful, ethical and responsible AI use. ........................................................................................................................................... 71 C77: Embed ethical principles in AI decision-making and operations. ............................... 72 C78: Select AI vendors and suppliers through fair and evidence-based evaluation. ......... 73 C79: Select technology platforms that meet organisational needs and mitigate long-term risks. ................................................................................................................................... 73 C80: Plan and manage workforce transformation resulting from AI adoption. ................... 74 C81: Drive reskilling and learning to prepare staff for AI-enabled roles. ............................ 75 C82: Manage AI risks within the organisation’s broader enterprise risk framework. .......... 76 C83: Ensure readiness for AI-related audits and incidents. ............................................... 77 C84: Create a culture that enables responsible innovation in AI. ...................................... 77 C85: Build trust and transparency in organisational AI use. ............................................... 78 C86: Act as a credible advocate for the organisation in AI policy and regulatory discussions. ........................................................................................................................................... 79 C87: Build trust with the public and external communities on AI use. ................................ 80 C88: Monitor regulatory changes and prepare the organisation for compliance. ............... 81
Consultation Draft for Review (December 2025). 7 C89: Oversee the embedding of regulatory readiness into business processes. .............. 81 C90: Supports peers and communities to understand and use AI responsibly through clear guidance and examples. .................................................................................................... 82 C91: Shares AI knowledge and practices in ways that are understandable to colleagues of different backgrounds. ........................................................................................................ 82 C92: Helps others understand their data and consumer rights and encourages the use of service settings that prevent unnecessary sharing of user data or contribution of content to model training. .................................................................................................................... 83 Annex A: AI Project Lifecycle Stages ..................................................................................... 85 Problem Definition .............................................................................................................. 86 Automation Discovery ........................................................................................................ 86 Requirements Gathering .................................................................................................... 87 Organisational Readiness Assessment .............................................................................. 87 Delivery Pathway Selection ................................................................................................ 87 Solution Design and Data Assessment .............................................................................. 88 Implementation Planning .................................................................................................... 88 Data Processing ................................................................................................................. 88 Solution Development ........................................................................................................ 88 Deployment and Implementation ........................................................................................ 89 Monitoring and Maintenance .............................................................................................. 89 Impact Evaluation ............................................................................................................... 89 System Retirement ............................................................................................................. 89 Governance ........................................................................................................................ 89 Risk Management .............................................................................................................. 90 Stakeholder Engagement ................................................................................................... 90
Consultation Draft for Review (December 2025). 8 Structure of the Framework The AI Skills for Business Competency Framework is designed as a practical, navigable guide to understanding the competencies required for the safe, effective and responsible use of AI across the workforce. The framework brings together a set of interlinked components: Personas, Duties, Competencies, Knowledge, Skill and Behaviour (KSB) statements, AI Project Lifecycle Stages and Proficiency Levels. Together, these describe not only what individuals need to be able to do, but why, when and to what depth. Personas help users identify the types of responsibilities they hold; Duties outline the tasks associated with those responsibilities. Competencies describe the capabilities needed to fulfil them, while Knowledge, Skill and Behaviour (KSB) statements break those competencies into their constituent parts. Meanwhile, Lifecycle Stages situate them within the broader process of AI development and use, and Proficiency Levels indicate the complexity and maturity at which they must be demonstrated. Personas Personas represent common patterns in how individuals engage with AI, but they are not intended to map directly onto job titles. Many roles span multiple types of responsibility, meaning an individual may simultaneously perform duties associated with Worker, Professional and Leadership personas. Users should therefore select and combine relevant personas to form a profile that accurately reflects the demands of their role. Duties For each persona, the framework identifies an example set of Duties. These duties capture common tasks and responsibilities associated with the given persona and provide a clear description of what individuals are expected to do. Duty statements offer clear, accessible language that individuals can use to understand expectations of their role. Each duty is underpinned by a series of competencies required to carry it out effectively. Competencies Competencies outline the specific capabilities an individual must demonstrate to perform duties effectively. Competencies are supported by the knowledge required to successfully carry out a duty, the skills they must be able to undertake, and the behaviours and values which underpin safe and responsible AI practices. We adopt the Skills England definitions of knowledge, skills and behaviours1 as follows: - Knowledge: “Knowledge is the information, technical detail, and ‘know-how’ that someone needs to have and understand to successfully carry out the duties.” - Skills: “Skills are the practical application of knowledge needed to successfully undertake the duties. They are learnt through training or experience.” - Behaviours: “Behaviours are mindsets, attitudes or approaches needed for competence. They may be more similar across occupations than knowledge and skills.” Knowledge, Skills and Behaviour (KSB) statements play a crucial role in operationalising competencies. They break competencies into clear units, helping to ensure competencies 1 Skills England (Accessed 2025). What is an occupa;onal standard?. h>ps://occupa;onalmaps.skillsengland.educa;on.gov.uk/what-is-an-occupa;onal-standard/
Consultation Draft for Review (December 2025). 9 are interpreted consistently. They support more precise identification of training skills needs, and support more equitable assessment of one's proficiency. They provide a bridge between competencies and the practical expectations for individuals to put these into practice. Proficiency Levels Proficiency denotes the level of mastery an individual exhibits in applying knowledge and skills within a domain. We draw on Anderson and Krathwohl’s (2021) taxonomy2, itself based on Bloom’s, to classify the type of knowledge a user engages with, and the cognitive processes they apply in their work. We extend Anderson and Krathwohl’s Cognitive Process Dimensions to include subcategories, demonstrating where this has a particular focus on values, professional development, leadership or accountability. This enables us to capture a progression from foundational understanding of key AI concepts, through to the ability to use this knowledge in evaluative and creative ways. Cognitive Process Dimension Description Example Level 1: Remember Recall facts, terminology or principles. Recalls definitions of fairness, bias and accountability. Level 2: Understand Explain concepts and interpret meaning. Describes ethical trade-offs in data collection practices. Level 3: Apply Use procedures or methods appropriately. Applies ethical principles when evaluating datasets. Level 3: Apply (L) Subcategory concerned with professional development. Facilitating learning among peers and across the organisation. Level 3: Apply (LA) Subcategory concerned with leadership. Setting direction, encouraging others, ensuring accountability, taking ownership of decisions. Level 3: Apply (C) Subcategory concerned with stakeholder engagement and collaboration. Ensuring effective communication, sharing of goals and coordinated efforts. Level 3: Apply (V) Subcategory concerned with individual professional values. Enabling trustworthy collaboration across the organisation. Level 4: Analyse Deconstruct problems, identify relationships. Identifies sources of bias in model development workflows. Level 5: Evaluate Judge quality, justify choices, and critique practices. Evaluates governance processes for ethical compliance. Level 6: Create Synthesise new ideas, innovate, or design frameworks. Designs new ethical guidance or audit frameworks for AI systems. We assign each Knowledge, Skills and Behaviour (KSB) statement a proficiency level according to the classification above. The complexity/proficiency of a given competency is determined by the highest level of any given KSB associated with that competency. By 2 Anderson, L. W. et al. (2014). A taxonomy for learning, teaching, and assessing: A revision of Bloom’s (Pearson new interna;onal ed.). Pearson.
Consultation Draft for Review (December 2025). 16 – D101: Act as a trusted Change Champion by supporting colleagues to adopt AI responsibly, modelling good practice, escalating risks, and helping teams navigate the human and practical impacts of AI-enabled change. – C26: Serve as a local advocate for the effective and ethical use of AI tools and practices within their team or department. – C27: Provide first-line guidance and support to co-workers on appropriate AI use and troubleshooting common issues. – C28: Encourage compliance with organisational AI policies and data governance standards, escalating risks or concerns to management when necessary. – C29: Identify areas where AI can support efficiency, innovation, or service improvements, and share recommendations with leadership. – C30: Encourage participation in AI training, sharing resources and best practice examples with colleagues. – C31: Act as a liaison between frontline staff and senior leadership, feeding back on staff needs, adoption challenges, and success stories. – C32: Keep up to date with developments in AI tools and organisational policy, ensuring colleagues are aware of relevant updates and guidance. – C33: Demonstrate effective use of AI by modelling good practice of AI use in dayto-day tasks, showcasing practical and responsible applications that inspire peers. Professional Professionals contribute to shaping how AI is built, evaluated and used. Their primary responsibilities involve the design, development, deployment, governance or strategic application of AI systems and tooling. This includes a wide range of roles across technical, research, policy, ethical, management and operational domains. They understand the scope and impact of their work within the broader AI lifecycle, including how it interacts with issues of safety, fairness, transparency and accountability. Each professional is aware of their responsibilities in relation to the systems they influence, be it through engineering, oversight, integration, or advisory functions. AI Professionals work with others across disciplines to support responsible, trustworthy AI. As stewards of capability, AI professionals play a critical role in ensuring that AI systems are technically robust, socially attuned and aligned with relevant policy. – D12: Contribute to the design, development, testing, or maintenance of AI systems to meet organisational or sector needs, ensuring they are technically robust, ethically sound, and aligned with intended outcomes. – C35: Design and develop data-driven solutions and AI systems. – C36: Test and monitor AI systems to assess performance and accuracy. – C37: Ensure that AI systems are designed and developed in an ethical, safe and responsible way. – C38: Project and change management through delivery of innovative projects. – C39: Apply rigorous methodology to generate, validate and share knowledge that supports responsible AI practices. – C40: Design and maintains data architectures, pipelines, and infrastructure that support reliable, secure, and scalable AI operations. – C41: Evaluate human-AI collaboration and ensure AI systems are aligned with user needs. – C42: Contribute to the design, development and deployment of AI systems in alignment with defined purpose, policy and ethical standards that include safety, robustness and reliability. – C43: Uphold professional standards in data stewardship, model evaluation and
Consultation Draft for Review (December 2025). 17 stakeholder engagement. – C44: Design AI architectures that align technical, ethical, and organisational objectives. – C45: Effectively select, integrate, and contribute to open-source tools, frameworks, and libraries that support robust and reproducible AI system development. – C46: Engage with open-source communities in ways that uphold ethical, secure, and inclusive practices, promoting responsible participation in the AI ecosystem. – C47: Collect and prepare the required datasets in line with organisation standards, ensuring accuracy, compliance and readiness for AI project use. – C53: Design and operate AI systems with due consideration for environmental and social sustainability. – C49: Collaborate effectively in multidisciplinary AI development teams. – C50: Implement, customise, and optimise low-code and no-code solutions that address organisational challenges, improve workflows, and enhance efficiency. – D13: Ensure AI solutions comply with legal, regulatory, and ethical standards, including data protection and fairness requirements. – C37: Ensure that AI systems are designed and developed in an ethical, safe and responsible way. – C51: Advocate for and ensure automation solutions are implemented responsibly, with fairness, transparency, and consideration for workforce wellbeing. – C52: Develop and apply methods to ensure AI explainability and interpretability to ensure that AI systems are transparent, interpretable, and understandable to appropriate audiences. – C42: Contribute to the design, development and deployment of AI systems in alignment with defined purpose, policy and ethical standards that include safety, robustness and reliability. – C48: Support the application of AI and related technologies in ways that uphold rights, promote trust and deliver value. – C47: Collect and prepare the required datasets in line with organisation standards, ensuring accuracy, compliance and readiness for AI project use. – C53: Design and operate AI systems with due consideration for environmental and social sustainability. – C49: Collaborate effectively in multidisciplinary AI development teams. – C54: Apply AI governance and risk management principles to ensure compliance and accountability. – C55: Support the development of AI systems to proactively manage potential workforce impacts. – C56: Support the responsible, ethical and safe adoption of AI. – D14: Produce clear documentation, audit trails, and model cards to enable transparency and accountability. – C57: Ensure responsible governance, security and stewardship of data across the AI lifecycle. – C58: Support documentation and auditability of AI systems and tooling. – D15: Monitor system performance, identifying and resolving risks such as bias, security vulnerabilities, or degraded accuracy. – C36: Test and monitor AI systems to assess performance and accuracy. – C37: Ensure that AI systems are designed and developed in an ethical, safe and responsible way. – C59: Deploy data-driven and AI solutions and integrate them with organisation's
Consultation Draft for Review (December 2025). 18 systems. – C39: Apply rigorous methodology to generate, validate and share knowledge that supports responsible AI practices. – C60: Maintain and continuously improve deployed AI systems. – C61: Ensure the security, robustness, and resilience of AI systems. – C40: Design and maintains data architectures, pipelines, and infrastructure that support reliable, secure, and scalable AI operations. – C58: Support documentation and auditability of AI systems and tooling. – C43: Uphold professional standards in data stewardship, model evaluation and stakeholder engagement. – C62: Contribute to the ongoing evaluation and refinement of solutions to ensure they remain effective and responsible, and aligned with evolving contexts. – C63: Performance management and continuous improvement. – D16: Explain technical concepts and uncertainties to business process owners, including managers and regulators. – C64: Engage senior leadership, end-users and relevant stakeholders across all phases of the Data and AI Lifecycle to ensure systems and data-driven solutions are fit for purpose. – C52: Develop and apply methods to ensure AI explainability and interpretability to ensure that AI systems are transparent, interpretable, and understandable to appropriate audiences. – C65: Support leadership and advocate for responsible introduction of artificial intelligence solutions. – C66: Communicate effectively with a range of audiences, ensuring clarity, transparency, and timeliness in reporting progress on AI throughout the AI project lifecycle. – D17: Collaborate across disciplines to align AI solutions with business goals, user requirements, and contextually relevant societal considerations. – C64: Engage senior leadership, end-users and relevant stakeholders across all phases of the Data and AI Lifecycle to ensure systems and data-driven solutions are fit for purpose. – C67: Design and develop data-driven and AI solutions that are intuitive and human-centred. – C68: Collaborate across disciplines to identify risks, integrate diverse perspectives and support responsible innovation. – C65: Support leadership and advocate for responsible introduction of artificial intelligence solutions. – C53: Design and operate AI systems with due consideration for environmental and social sustainability. – C66: Communicate effectively with a range of audiences, ensuring clarity, transparency, and timeliness in reporting progress on AI throughout the AI project lifecycle. – D18: Keep technical knowledge current and supports continuous learning for themselves and others. – C34: Demonstrate ownership of personal development and engage in continuous learning activities. – C69: Research, critically evaluate, and apply emerging knowledge to AI automation developments, ensuring practice remains current, ethical, and forward-looking. – C32: Keep up to date with developments in AI tools and organisational policy,
Consultation Draft for Review (December 2025). 19 ensuring colleagues are aware of relevant updates and guidance. – C70: Evaluate emerging AI technologies and their potential societal impact. – C71: Support learning, mentorship, and capacity building in responsible and ethical AI practice. – C62: Contribute to the ongoing evaluation and refinement of solutions to ensure they remain effective and responsible, and aligned with evolving contexts. – D100: Support the responsible evaluation and ongoing oversight of external AI vendors and third-party tools. – C72: Engage with and/or develop governance processes that enable validation, oversight and escalation. – C73: Evaluate and oversee external AI vendors and third-party tools to ensure technical fitness, compliance, and responsible operation. Leader Leaders are responsible for setting the strategic direction and governance conditions for how AI is adopted and used within an organisation or industry. They may hold executive, programme or policy level roles and are accountable for ensuring that, the regulation compliant, AI technologies they select, align with organisational goals, values and long-term public trust. While not expected to be technical experts, AI Leaders are literate in the general capabilities, limitations, and governance requirements of the AI systems they choose to adopt. This enables them to make informed decisions in complex and evolving environments. They can challenge AI offerings that pose ethical, legal, reputational or operational risks by asking the right questions and setting clear guardrails to ensure that AI integration reflects the organisation’s risk appetite and strategic priorities. They champion responsible innovation and build cross-functional collaborative teams and understand the impact of AI adoption on their workforce. They support staff through organisational change and invest in the right skills to enable confident and accountable use of AI. – D19: Set the organisation’s overall strategy for AI adoption and ensures it aligns with business objectives and values. – C74: Define and communicate a coherent AI strategy that aligns with organisational goals and values. – C75: Balance innovation with practical delivery in AI adoption. – D20: Provide governance and oversight of AI use, ensuring compliance with laws, regulations, and ethical principles. – C76: Sponsor governance frameworks that safeguard lawful, ethical and responsible AI use. – C77: Embed ethical principles in AI decision-making and operations. – D21: Make procurement and investment decisions about AI tools, platforms, and external suppliers. – C78: Select AI vendors and suppliers through fair and evidence-based evaluation. – C79: Select technology platforms that meet organisational needs and mitigate long-term risks. – D22: Anticipate workforce impacts of AI and invest in upskilling, reskilling, and change management.
Consultation Draft for Review (December 2025). 20 – C80: Plan and manage workforce transformation resulting from AI adoption. – C81: Drive reskilling and learning to prepare staff for AI-enabled roles. – D23: Oversee risk management, ensuring organisational readiness for audits, assurance, and incident response. – C83: Ensure readiness for AI-related audits and incidents. – C82: Manage AI risks within the organisation’s broader enterprise risk framework. – D24: Promote a culture of responsible AI use, encouraging innovation within safe and transparent boundaries. – C84: Create a culture that enables responsible innovation in AI. – C85: Build trust and transparency in organisational AI use. – D25: Represent the organisation externally in discussions with regulators, partners, and the public on AI issues. – C86: Act as a credible advocate for the organisation in AI policy and regulatory discussions. – C87: Build trust with the public and external communities on AI use. – D26: Anticipate and drive the organisation's preparedness for future AI regulation. – C88: Monitor regulatory changes and prepare the organisation for compliance. – C89: Oversee the embedding of regulatory readiness into business processes. Citizen Advocate AI Citizen Advocates act as trusted enablers within their communities, helping others to understand and engage confidently with artificial intelligence in everyday life. They might lead discussions in local settings such as libraries, community centres, or schools, supporting others to recognise where AI is used and how it impacts daily activities. While they are not technical experts, AI Citizen Advocates promote responsible and inclusive dialogue about AI’s benefits, risks, and limitations. They help others access trustworthy information, challenge misinformation, and make informed choices about data sharing and AI-generated content. Through empathy, curiosity, and clear communication, AI Citizen Advocates build community confidence in navigating an AI-enabled world. They promote critical thinking, ethical awareness, and participation in shaping how AI technologies serve people and society. – D28: Support others to recognise where AI is used in daily life and understand its benefits, risks, and limitations. – C90: Supports peers and communities to understand and use AI responsibly through clear guidance and examples. – C91: Shares AI knowledge and practices in ways that are understandable to colleagues of different backgrounds. – D29: Promote responsible and informed use of AI-enabled tools within communities. – C90: Supports peers and communities to understand and use AI responsibly through clear guidance and examples. – C92: Helps others understand their data and consumer rights and encourages the use of service settings that prevent unnecessary sharing of user data or contribution of content to model training.
Consultation Draft for Review (December 2025). 21 – D30: Help others recognise and respond to AI-generated misinformation, bias, or harm. – C11: Recognises and responds to instances of unfairness, exclusion, or bias in AI systems and promotes equitable practices. – C12: Knows where and how to report AI-related scams, abuse, or unfair outcomes and helps others do the same. – D31: Build community confidence and inclusion in AI discussions and decision-making. – C14: Contributes to AI related discussions by actively listening, sharing perspectives and respecting diverse experience. – C15: Makes space for different experiences and needs so public conversations about AI are fair and inclusive. – D32: Connect communities with reliable information, learning opportunities, and support on AI. – C90: Supports peers and communities to understand and use AI responsibly through clear guidance and examples. – C92: Helps others understand their data and consumer rights and encourages the use of service settings that prevent unnecessary sharing of user data or contribution of content to model training. – D33: Champion ethical awareness and responsible AI citizenship. – C11: Recognises and responds to instances of unfairness, exclusion, or bias in AI systems and promotes equitable practices. – C13: Contributes to conversations about AI with respect and evidence. – D34: Continuously develop understanding of AI's societal effects and share learning with others. – C13: Contributes to conversations about AI with respect and evidence. – C91: Shares AI knowledge and practices in ways that are understandable to colleagues of different backgrounds.
Consultation Draft for Review (December 2025). 22 Summary of Competencies This section sets out the full list of competencies included in the framework. Subsequent sections detail each competency in full, describing the knowledge, skills, and behaviours required to carry out the associated duties effectively. C1: Identifies where AI operates in daily life and understands its basic functions and limits. C2: Engages with AI-driven tools and services thoughtfully, understanding both benefits and risks. C3: Understands user rights, choices and complaint routes when AI affects access, pricing, or critical services and opportunities. C4: Understands how personal data is collected, processed, and used to train and personalise AI systems. C5: Takes active steps to protect personal data and manage privacy across AI-enabled platforms. C6: Shares only what is necessary, understands consent choices, and reviews permissions regularly. C7: Demonstrates awareness of AI-generated material and understands its implications for information integrity. C8: Critically assesses the reliability and fairness of AI-generated outputs before accepting or sharing them. C9: Demonstrates awareness of the risks of sharing AI-generated content by pausing, checking context, and adding sources when appropriate. C10: Detects and avoids harmful or deceptive uses of AI, including fraud, manipulation, or misinformation. C11: Recognises and responds to instances of unfairness, exclusion, or bias in AI systems and promotes equitable practices. C12: Knows where and how to report AI-related scams, abuse, or unfair outcomes and helps others do the same. C13: Contributes to conversations about AI with respect and evidence. C14: Contributes to AI related discussions by actively listening, sharing perspectives and respecting diverse experience. C15: Makes space for different experiences and needs so public conversations about AI are fair and inclusive. C16: Comply with data protection policies and safeguard sensitive information when using AI systems. C17: Use AI-powered platforms confidently and effectively in everyday work. C18: Explore and experiment with AI tools responsibly, ensuring alignment with organisational policies. C19: Demonstrate and promote ethical awareness in everyday AI use by respecting privacy, questioning outputs, and encouraging responsible practices among colleagues. C20: Use AI safely and inclusively in collaborative settings, promoting transparency and equitable participation, with awareness of potential bias and accessibility limitations. C21: Be transparent about the role of AI in producing work outputs. C22: Apply AI tooling into workflows responsibly, ensuring alignment with organisational
Consultation Draft for Review (December 2025). 23 policies and practices. C23: Interpret AI-driven outputs critically and apply professional judgement. C24: Adapt to the changing nature of work through ongoing engagement with AI tools. C25: Contribute to organisational learning about AI use. C26: Serve as a local advocate for the effective and ethical use of AI tools and practices within their team or department. C27: Provide first-line guidance and support to co-workers on appropriate AI use and troubleshooting common issues. C28: Encourage compliance with organisational AI policies and data governance standards, escalating risks or concerns to management when necessary. C29: Identify areas where AI can support efficiency, innovation, or service improvements, and share recommendations with leadership. C30: Encourage participation in AI training, sharing resources and best practice examples with colleagues. C31: Act as a liaison between frontline staff and senior leadership, feeding back on staff needs, adoption challenges, and success stories. C32: Keep up to date with developments in AI tools and organisational policy, ensuring colleagues are aware of relevant updates and guidance. C33: Demonstrate effective use of AI by modelling good practice of AI use in day-to-day tasks, showcasing practical and responsible applications that inspire peers. C34: Demonstrate ownership of personal development and engage in continuous learning activities. C35: Design and develop data-driven solutions and AI systems. C36: Test and monitor AI systems to assess performance and accuracy. C37: Ensure that AI systems are designed and developed in an ethical, safe and responsible way. C38: Project and change management through delivery of innovative projects. C39: Apply rigorous methodology to generate, validate and share knowledge that supports responsible AI practices. C40: Design and maintains data architectures, pipelines, and infrastructure that support reliable, secure, and scalable AI operations. C41: Evaluate human-AI collaboration and ensure AI systems are aligned with user needs. C42: Contribute to the design, development and deployment of AI systems in alignment with defined purpose, policy and ethical standards that include safety, robustness and reliability. C43: Uphold professional standards in data stewardship, model evaluation and stakeholder engagement. C44: Design AI architectures that align technical, ethical, and organisational objectives. C45: Effectively select, integrate, and contribute to open-source tools, frameworks, and libraries that support robust and reproducible AI system development. C46: Engage with open-source communities in ways that uphold ethical, secure, and inclusive practices, promoting responsible participation in the AI ecosystem. C47: Collect and prepare the required datasets in line with organisation standards, ensuring accuracy, compliance and readiness for AI project use.
Consultation Draft for Review (December 2025). 24 C48: Support the application of AI and related technologies in ways that uphold rights, promote trust and deliver value. C49: Collaborate effectively in multidisciplinary AI development teams. C50: Implement, customise, and optimise low-code and no-code solutions that address organisational challenges, improve workflows, and enhance efficiency. C51: Advocate for and ensure automation solutions are implemented responsibly, with fairness, transparency, and consideration for workforce wellbeing. C52: Develop and apply methods to ensure AI explainability and interpretability to ensure that AI systems are transparent, interpretable, and understandable to appropriate audiences. C53: Design and operate AI systems with due consideration for environmental and social sustainability. C54: Apply AI governance and risk management principles to ensure compliance and accountability. C55: Support the development of AI systems to proactively manage potential workforce impacts. C56: Support the responsible, ethical and safe adoption of AI. C57: Ensure responsible governance, security and stewardship of data across the AI lifecycle. C58: Support documentation and auditability of AI systems and tooling. C59: Deploy data-driven and AI solutions and integrate them with organisation's systems. C60: Maintain and continuously improve deployed AI systems. C61: Ensure the security, robustness, and resilience of AI systems. C62: Contribute to the ongoing evaluation and refinement of solutions to ensure they remain effective and responsible, and aligned with evolving contexts. C63: Performance management and continuous improvement. C64: Engage senior leadership, end-users and relevant stakeholders across all phases of the Data and AI Lifecycle to ensure systems and data-driven solutions are fit for purpose. C65: Support leadership and advocate for responsible introduction of artificial intelligence solutions. C66: Communicate effectively with a range of audiences, ensuring clarity, transparency, and timeliness in reporting progress on AI throughout the AI project lifecycle. C67: Design and develop data-driven and AI solutions that are intuitive and human-centred. C68: Collaborate across disciplines to identify risks, integrate diverse perspectives and support responsible innovation. C69: Research, critically evaluate, and apply emerging knowledge to AI automation developments, ensuring practice remains current, ethical, and forward-looking. C70: Evaluate emerging AI technologies and their potential societal impact. C71: Support learning, mentorship, and capacity building in responsible and ethical AI practice. C72: Engage with and/or develop governance processes that enable validation, oversight and escalation. C73: Evaluate and oversee external AI vendors and third-party tools to ensure technical fitness, compliance, and responsible operation.
Consultation Draft for Review (December 2025). 25 C74: Define and communicate a coherent AI strategy that aligns with organisational goals and values. C75: Balance innovation with practical delivery in AI adoption. C76: Sponsor governance frameworks that safeguard lawful, ethical and responsible AI use. C77: Embed ethical principles in AI decision-making and operations. C78: Select AI vendors and suppliers through fair and evidence-based evaluation. C79: Select technology platforms that meet organisational needs and mitigate long-term risks. C80: Plan and manage workforce transformation resulting from AI adoption. C81: Drive reskilling and learning to prepare staff for AI-enabled roles. C82: Manage AI risks within the organisation’s broader enterprise risk framework. C83: Ensure readiness for AI-related audits and incidents. C84: Create a culture that enables responsible innovation in AI. C85: Build trust and transparency in organisational AI use. C86: Act as a credible advocate for the organisation in AI policy and regulatory discussions. C87: Build trust with the public and external communities on AI use. C88: Monitor regulatory changes and prepare the organisation for compliance. C89: Oversee the embedding of regulatory readiness into business processes. C90: Supports peers and communities to understand and use AI responsibly through clear guidance and examples. C91: Shares AI knowledge and practices in ways that are understandable to colleagues of different backgrounds. C92: Helps others understand their data and consumer rights and encourages the use of service settings that prevent unnecessary sharing of user data or contribution of content to model training.
Consultation Draft for Review (December 2025). 32 information when hosting or joining discussions. (3 Apply - C) Behaviour Statements: – B97: Shows respect in disagreement and actively includes people with different backgrounds or views. (3 Apply - V) C16: Comply with data protection policies and safeguard sensitive information when using AI systems. Proficiency Level: 5 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K94: Understands organisational data protection policies and broader legal frameworks (e.g. GDPR) and why compliance is critical. (2 Understand) – K95: Knows what constitutes sensitive or business-critical information, including Personal identifiable information (PII), financial records, or confidential project material. (2 Understand) – K96: Understands the risks of entering sensitive information into AI systems, and the distinction between approved and unapproved tools. (2 Understand) Skill Statements: – S178: Can correctly identify sensitive information and manage it according to policy when using AI tools. (5 Evaluate) – S179: Demonstrates the ability to apply secure practices, such as anonymisation, restricted sharing, and safe storage. (3 Apply) – S180: Can decide whether a task should be performed using internal organisational systems or whether the use of external platforms is appropriate and compliant. (5 Evaluate) Behaviour Statements: – B47: Acts with care, responsibility, and integrity in handling data. (3 Apply - V) – B48: Consistently applies organisational rules, avoiding shortcuts that could compromise security. (3 Apply - V) – B49: Exercises caution with unapproved systems and defaults to safe, policy-compliant tools. (3 Apply - V) C17: Use AI-powered platforms confidently and effectively in everyday work. Proficiency Level: 6 Lifecycle Stage(s): Build and Deploy | Operate and Evaluate Knowledge Statements: – K97: Understands the basic capabilities and limitations of AI-powered workplace tools (e.g. productivity, HR, finance, or customer service platforms). (2 Understand) – K98: Recognises that AI tools are assistive, not authoritative, and that human judgement remains essential. (4 Analyse) – K99: Knows how to seek help or escalate issues if a tool produces unexpected, unclear, or unsafe outputs. (5 Evaluate) Skill Statements: – S181: Can use AI-powered platforms competently to support everyday tasks, such as document drafting, scheduling, data analysis, or content creation. (3 Apply) – S182: Demonstrates proficiency in operating AI tools, navigating interfaces, applying key features, and resolving common issues independently. (3 Apply)
Consultation Draft for Review (December 2025). 33 – S183: Can use AI tools with other digital platforms and workflows to enhance productivity, automate routine tasks, and achieve efficient outcomes. (6 Create) Behaviour Statements: – B50: Approaches AI systems with confidence, curiosity, and a willingness to learn. (3 Apply - L) – B51: Demonstrates adaptability when platforms are implemented, updated or expanded. (3 Apply - L) – B52: Uses AI responsibly, never over-relying on outputs without critical review. (5 Evaluate) C18: Explore and experiment with AI tools responsibly, ensuring alignment with organisational policies. Proficiency Level: 4 Lifecycle Stage(s): Build and Deploy | Operate and Evaluate Knowledge Statements: – K100: Understands that AI systems are continually evolving, with outputs that may vary in quality and reliability. (2 Understand) Skill Statements: – S184: Can try out approved AI tools for appropriate tasks, experimenting with prompts and features to understand their potential. (3 Apply) – S185: Demonstrates the ability to capture and reflect on lessons from experimentation. (4 Analyse) – S187: Can share insights with colleagues to promote organisational learning. (3 Apply - L) Behaviour Statements: – B54: Demonstrates curiosity and openness, approaching AI with a learner mindset. (3 Apply - L) – B55: Acts responsibly, avoiding experimentation that could expose sensitive data or create reputational risk. (3 Apply - V) – B56: Recognises that lessons learned from AI tool exploration should be documented and shared where appropriate, to support transparency, peer learning, and responsible practice. (3 Apply - C) C19: Demonstrate and promote ethical awareness in everyday AI use by respecting privacy, questioning outputs, and encouraging responsible practices among colleagues. Proficiency Level: 5 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K109: Understands key ethical considerations in AI, including bias, fairness, privacy, inclusivity, and accountability. (2 Understand) – K110: Recognises ethical risks in routine use, such as misuse of outputs, inappropriate data entry, or perpetuating biases. (4 Analyse) – K111: Knows how to report or escalate ethical concerns. (2 Understand) Skill Statements: – S194: Can identify and respond to ethical concerns in AI use. (5 Evaluate) – S195: Demonstrates the ability to use AI responsibly, respecting privacy, inclusivity, and
Consultation Draft for Review (December 2025). 34 organisational values. (3 Apply - LA) – S196: Can raise ethical issues constructively with managers, technical teams, or ethics leads. (3 Apply - LA) Behaviour Statements: – B63: Acts with integrity when using AI tools, even for routine tasks. (3 Apply - V) – B65: Promotes ethical practice by encouraging colleagues to think critically about AI use. (3 Apply - LA) – B66: Raises concerns when AI use creates risks to fairness, privacy, or appropriate work practices. (3 Apply - LA) C20: Use AI safely and inclusively in collaborative settings, promoting transparency and equitable participation, with awareness of potential bias and accessibility limitations. Proficiency Level: 5 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K112: Understands risks of using AI in collaboration (e.g. transcription tools, shared documents, or meeting summaries), including issues of consent, privacy, bias and errors. (2 Understand) – K114: Knows organisational policies for responsible use of AI in shared spaces. (2 Understand) – K115: Recognises the need for inclusivity, ensuring the use of AI does not disadvantage colleagues with different voices, accents, or roles. (4 Analyse) Skill Statements: – S198: Can apply safe practices when using AI in meetings or team contexts, including seeking consent and reviewing outputs before sharing. (3 Apply) – S199: Demonstrates the ability to identify and report inclusivity or accessibility risks in AI systems and workflows. (5 Evaluate) – S200: Can provide feedback on collaboration tools to improve accuracy and fairness. (5 Evaluate) Behaviour Statements: – B67: Demonstrates respect for colleagues’ privacy and preferences. (3 Apply - V) – B68: Acts inclusively, ensuring AI use supports rather than excludes participants. (3 Apply - V) – B69: Takes responsibility for validating and correcting shared AI outputs. (3 Apply - LA) C21: Be transparent about the role of AI in producing work outputs. Proficiency Level: 5 Lifecycle Stage(s): Build and Deploy | Operate and Evaluate Knowledge Statements: – K116: Understands organisational expectations around transparency in AI use, including when attribution is required. (2 Understand) – K117: Knows why transparency matters for accountability, trust, and professional credibility. (2 Understand) – K118: Recognises contexts where disclosure is essential, especially for shared, publicfacing, or critical outputs. (4 Analyse)
Consultation Draft for Review (December 2025). 35 Skill Statements: – S201: Can acknowledge AI contributions when outputs are shared internally or externally. (3 Apply) – S202: Demonstrates awareness of the challenges in reliably distinguishing between human-authored and AI-assisted work, and applies critical judgement when assessing the likely origin of content. (5 Evaluate) – S203: Can apply organisational attribution guidelines consistently. (3 Apply) Behaviour Statements: – B70: Models openness around disclosing use of AI, encouraging colleagues to follow suit. (3 Apply - LA) C22: Apply AI tooling into workflows responsibly, ensuring alignment with organisational policies and practices. Proficiency Level: 6 Lifecycle Stage(s): Build and Deploy | Operate and Evaluate Knowledge Statements: – K103: Understands common patterns of human-AI collaboration, including automation, decision support, and content generation. (2 Understand) – K104: Knows how to judge when AI use is appropriate, and when human oversight or a fully human-led process is required. (4 Analyse) – K105: Recognises the importance of escalation to managers or technical teams where AI results raise risks or uncertainties. (2 Understand) Skill Statements: – S188: Can combine AI outputs with human expertise to improve efficiency and quality of work. (6 Create) – S189: Uses AI-enabled tools appropriately within existing workflows, following guidance to complete tasks safely, efficiently, and responsibly (6 Create) – S190: Applies critical analysis to determine when human oversight or specialist intervention is required instead of AI. (4 Analyse) Behaviour Statements: – B57: Balances efficiency gains with quality and accountability. (3 Apply - LA) – B58: Remains open to collaboration with AI systems while maintaining human responsibility. (3 Apply - LA) – B59: Demonstrates initiative by identifying opportunities for workflow improvement. (4 Analyse) C23: Interpret AI-driven outputs critically and apply professional judgement. Proficiency Level: 5 Lifecycle Stage(s): Build and Deploy | Operate and Evaluate Knowledge Statements: – K106: Understands that AI outputs may be inaccurate, biased, or incomplete. (4 Analyse) – K107: Knows the importance of cross-checking outputs against trusted sources, policies, and expertise. (2 Understand) – K108: Recognises that responsibility for decisions based on AI outputs remains with the human worker. (4 Analyse)
Consultation Draft for Review (December 2025). 36 Skill Statements: – S191: Can evaluate AI outputs for accuracy, relevance, fairness, and appropriateness. (5 Evaluate) – S192: Demonstrates the ability to identify when outputs require correction, supplementation, or rejection. (5 Evaluate) – S193: Can escalate questionable outputs to supervisors or colleagues for validation. (3 Apply - LA) Behaviour Statements: – B60: Demonstrates accountability for AI-assisted work, never passing unchecked outputs as final. (3 Apply - LA) – B61: Shows critical thinking, avoiding blind reliance or indiscriminate acceptance of AI outputs. (5 Evaluate) – B62: Ensures professional standards of accuracy and fairness are upheld. (3 Apply - LA) C24: Adapt to the changing nature of work through ongoing engagement with AI tools. Proficiency Level: 4 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K119: Understands that AI adoption may change job roles, responsibilities, and required skills. (2 Understand) – K120: Knows how AI can both remove routine tasks and create new opportunities. (2 Understand) – K122: Recognises limitations of their work, and the importance of continuous learning and adaptation. (3 Apply - L) Skill Statements: – S204: Can reflect on how AI is affecting their role and respond to adapting workflows accordingly. (4 Analyse) – S206: Demonstrates the ability to highlight opportunities for efficiency or improvement. (4 Analyse) – S207: Can contribute constructively to discussions about workforce transformation and reskilling. (3 Apply - L) Behaviour Statements: – B72: Demonstrates adaptability and openness to role change. (3 Apply - V) – B73: Engages proactively in learning and skill development. (3 Apply - L) – B74: Shows responsibility by reporting risks and sharing opportunities created by AI in their work. (3 Apply - LA) C25: Contribute to organisational learning about AI use. Proficiency Level: 4 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K123: Understands that responsible AI adoption is an evolving practice requiring feedback and shared learning. (2 Understand) – K124: Recognises the importance of collective adaptation, beyond individual use. (4 Analyse) – K125: Knows how to provide input into organisational improvement processes. (2
Consultation Draft for Review (December 2025). 37 Understand) Skill Statements: – S208: Can share experiences, tips, and lessons learned from AI use with colleagues. (3 Apply - C) – S209: Demonstrates the ability to provide constructive feedback on tools and practices. (3 Apply - C) – S210: Can participate in reskilling or upskilling initiatives, contributing to organisational capability. (3 Apply - L) Behaviour Statements: – B75: Demonstrates collaboration by helping colleagues improve their confidence when using AI. (3 Apply - L) – B76: Acts constructively when raising concerns, focusing on solutions and shared improvement. (3 Apply - LA) – B77: Models continuous learning, showing that responsible AI use is a shared responsibility. (3 Apply - LA) C26: Serve as a local advocate for the effective and ethical use of AI tools and practices within their team or department. Proficiency Level: 3 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K201: Awareness of organisational AI policy, ethical guidelines, and relevant legislation. (2 Understand) – K203: Understanding of common AI applications and limitations in the workplace. (2 Understand) Skill Statements: – S280: Ability to explain AI concepts in clear, accessible language. (2 Understand) – S281: Ability to highlight risks and opportunities of AI use to colleagues. (2 Understand) Behaviour Statements: – B140: Acts as a positive role model for responsible AI use. (3 Apply - V) – B141: Demonstrates integrity when promoting AI solutions. (3 Apply - V) C27: Provide first-line guidance and support to co-workers on appropriate AI use and troubleshooting common issues. Proficiency Level: 6 Lifecycle Stage(s): Build and Deploy | Operate and Evaluate Knowledge Statements: – K244: Familiarity with organisationally approved AI tools and their functions. (2 Understand) – K245: Understanding of typical challenges and common mistakes in AI use. (2 Understand) – K246: Basic understanding of key AI concepts, including machine learning, natural language processing, and automation. (2 Understand) – K247: Awareness of common model limitations such as bias, hallucinations, and overfitting. (2 Understand) – K248: Understanding of common failure modes in AI use, including incorrect prompts, unrealistic expectations, or misuse of outputs. (2 Understand)
Consultation Draft for Review (December 2025). 38 – K249: Knowledge of basic diagnostic techniques for identifying and resolving common AI issues (e.g. reviewing logs, checking configurations, verifying data inputs). (2 Understand) – K250: Awareness of integration points between AI tools and wider organisational systems or data pipelines. (2 Understand) – K251: Understanding how to match AI tools and techniques to specific tasks and business needs. (2 Understand) – K252: Knowledge of organisational policies, data protection standards, and escalation routes for identifying and flagging risky or inappropriate AI use. (2 Understand) Skill Statements: – S325: Ability to provide practical, hands-on support to peers. (3 Apply - C) – S326: Strong interpersonal and communication skills to guide colleagues confidently. (3 Apply - C) – S327: Ability to diagnose and troubleshoot basic technical or configuration issues in AI tools. (5 Evaluate) – S328: Skill in interpreting and explaining model limitations and uncertainties in accessible language. (5 Evaluate) – S329: Capability to identify when AI outputs appear unreliable and to suggest corrective actions or escalation. (5 Evaluate) – S330: Ability to align AI tools and workflows with specific business or operational tasks to enhance productivity. (6 Create) – S331: Skill in recognising and addressing risky or non-compliant behaviours in AI use in line with organisational policy. (5 Evaluate) Behaviour Statements: – B142: Patient and approachable when supporting others. (3 Apply - C) – B143: Proactive in offering help and sharing expertise. (3 Apply - C) C28: Encourage compliance with organisational AI policies and data governance standards, escalating risks or concerns to management when necessary. Proficiency Level: 3 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K204: Awareness of data protection, information security, and AI governance requirements, where relevant. (2 Understand) – K205: Understanding of potential biases, risks, and misuse of AI. (2 Understand) Skill Statements: – S283: Ability to identify and escalate risks in AI use. (3 Apply - LA) – S284: Ability to reinforce compliance with organisational policies. (3 Apply - LA) Behaviour Statements: – B144: Vigilant in ensuring safe and ethical practice. (3 Apply - LA) – B145: Responsible in raising issues without blame, focusing on solutions. (3 Apply - LA) C29: Identify areas where AI can support efficiency, innovation, or service improvements, and share recommendations with leadership. Proficiency Level: 5 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate
Consultation Draft for Review (December 2025). 39 Knowledge Statements: – K206: Awareness of business processes and tasks where AI might add value. (2 Understand) – K207: Knowledge of current and emerging AI applications in the sector. (2 Understand) Skill Statements: – S285: Ability to spot inefficiencies and suggest AI-supported improvements. (5 Evaluate) – S286: Assesses opportunities for AI adoption by analysing workflows, data availability, and potential benefits e.g. innovation, or service quality. (3 Apply - L) Behaviour Statements: – B146: Demonstrates curiosity and a commitment to maintaining an awareness of AI developments through continued professional development. (3 Apply - L) – B147: Constructive in sharing recommendations with management. (3 Apply - C) C30: Encourage participation in AI training, sharing resources and best practice examples with colleagues. Proficiency Level: 3 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K208: Understanding of available AI training resources and learning pathways. (3 Apply - L) – K209: Awareness of colleagues’ varied digital literacy and skill levels. (3 Apply - L) Skill Statements: – S287: Ability to encourage and motivate peers to build confidence in AI. (3 Apply - LA) – S288: Ability to share learning materials and explain relevance to roles. (3 Apply - L) Behaviour Statements: – B146: Demonstrates curiosity and a commitment to maintaining an awareness of AI developments through continued professional development. (3 Apply - L) – B148: Supportive and inclusive, recognising different learning needs. (3 Apply - LA) C31: Act as a liaison between frontline staff and senior leadership, feeding back on staff needs, adoption challenges, and success stories. Proficiency Level: 3 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K210: Awareness of organisational AI strategy and leadership priorities. (2 Understand) – K212: Understanding of staff perspectives, adoption challenges, and needs. (2 Understand) Skill Statements: – S289: Ability to listen actively and capture feedback from colleagues. (3 Apply - C) – S290: Ability to present feedback clearly to leadership in a constructive way. (3 Apply - C) Behaviour Statements: – B147: Constructive in sharing recommendations with management. (3 Apply - C) – B150: Balanced and fair in representing views. (3 Apply - C)
Consultation Draft for Review (December 2025). 40 C32: Keep up to date with developments in AI tools and organisational policy, ensuring colleagues are aware of relevant updates and guidance. Proficiency Level: 3 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K213: Up-to-date awareness of AI developments and organisational updates to organisational policy. (3 Apply - L) – K214: Knowledge of credible information sources on AI. (1 Remember) Skill Statements: – S291: Ability to distil and share updates with colleagues effectively. (3 Apply - C) – S293: Ability to adapt practice in response to new guidance or tools. (3 Apply - L) Behaviour Statements: – B146: Demonstrates curiosity and a commitment to maintaining an awareness of AI developments through continued professional development. (3 Apply - L) – B151: Inspire others through openness and transparency, ensuring others are kept informed of developments, where relevant. (3 Apply - LA) C33: Demonstrate effective use of AI by modelling good practice of AI use in day-to-day tasks, showcasing practical and responsible applications that inspire peers. Proficiency Level: 3 Lifecycle Stage(s): Operate and Evaluate Knowledge Statements: – K215: Awareness of practical, safe, and productive uses of AI in the workplace. (2 Understand) – K216: Understanding of what constitutes poor or unsafe AI use. (2 Understand) Skill Statements: – S294: Ability to demonstrate AI tools in real work contexts. (3 Apply) – S295: Ability to embed AI responsibly in their own work tasks. (3 Apply - L) Behaviour Statements: – B140: Acts as a positive role model for responsible AI use. (3 Apply - V) – B151: Inspire others through openness and transparency, ensuring others are kept informed of developments, where relevant. (3 Apply - LA) C34: Demonstrate ownership of personal development and engage in continuous learning activities. Proficiency Level: 4 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K208: Understanding of available AI training resources and learning pathways. (3 Apply - L) – K169: Understanding of learning and development practices such as training, mentoring, and CPD. (2 Understand)
Consultation Draft for Review (December 2025). 41 Skill Statements: – S46: Learn from experience through self-assessment of one’s own responses to practice situations. (4 Analyse) – S47: Identify learning opportunities to maintain knowledge and skills in the relevant area of data science and AI (3 Apply) – S48: Engage with the latest developments across industry and academia and incorporate these into solutions. (3 Apply) – S1: Demonstrates a curious mindset and cultural intelligence around perceptions of datadriven and AI technology (3 Apply - L) – S1: Demonstrates a curious mindset and cultural intelligence around perceptions of datadriven and AI technology (3 Apply - L) Behaviour Statements: – B146: Demonstrates curiosity and a commitment to maintaining an awareness of AI developments through continued professional development. (3 Apply - L) – B113: Models continuous learning and openness to feedback. (3 Apply - L) – B1: Demonstrates openness to feedback and a willingness to adjust approaches based on evidence and colleague input. (3 Apply - V) C35: Design and develop data-driven solutions and AI systems. Proficiency Level: 6 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K191: Awareness of ethical principles and professional standards relevant to AI development (e.g., fairness, transparency, accountability, human dignity). (1 Remember) – K222: Understand the importance of designing AI systems that augment rather than unnecessarily replace human work where feasible. (2 Understand) – K33: Understanding of AI models architectures (2 Understand) – K34: Familiarity with supervised and unsupervised learning (2 Understand) – K37: Understands the importance of testing and monitoring AI systems to ensure reliability, fairness, safety, and compliance with organisational and regulatory standards. (2 Understand) – K38: Knows common testing and validation methods for AI systems, including crossvalidation, hold-out testing, A/B testing, and performance benchmarking against baseline models. (2 Understand) – K39: Knows how to define clear performance objectives and evaluation criteria to test AI systems against intended outcomes and success metrics. (2 Understand) – K40: Understands how testing and monitoring requirements vary across sectors, contexts, and risk levels, and how to tailor evaluation approaches accordingly. (4 Analyse) Skill Statements: – S57: Deliver secure, stable and scalable data products and AI solutions to meet the needs for the organisation (6 Create) – S58: Awareness of the benefits and limitations of open source software (2 Understand) – S59: Design and deliver data products and AI solutions that meet appropriate accessibility standards for their users (6 Create) – S60: Identify and elicit project requirements (5 Evaluate) – S61: Identify and quantify biases (5 Evaluate) – S62: Identifies, quantifies, and mitigates bias in data, models, or processes by using appropriate metrics, diagnostic tools, and intervention strategies to support fairness, accountability, and contextual relevance. (5 Evaluate)
Consultation Draft for Review (December 2025). 48 Apply - LA) Behaviour Statements: – B233: Demonstrates integrity, respect, and accountability in all aspects of professional conduct (2 Apply - V) – B234: Respects privacy, consent, and individual rights in data use (3 Apply - V) – B235: Builds and sustains trust with colleagues and stakeholders through transparency and reliability (3 Apply - LA) – B236: Models ethical behaviour and encourages others to uphold professional standards (3 Apply - LA) C44: Design AI architectures that align technical, ethical, and organisational objectives. Proficiency Level: 6 Lifecycle Stage(s): Design Knowledge Statements: – K322: Understands design principles for AI and data system architectures, including modularity, scalability, and maintainability. (2 Understand) – K323: Knows how to align technical design decisions with business objectives, user requirements, and ethical constraints. (2 Understand) – K324: Understands trade-offs between model complexity, explainability, energy efficiency, and fairness. (2 Understand) – K325: Knows key patterns and tools for managing data flows, APIs, and computational infrastructure across the AI lifecycle. (2 Understand) – K310: Knows the professional, legal, and societal responsibilities associated with AI and data use (2 Understand) Skill Statements: – S412: Designs and documents system architectures that integrate data, models, and interfaces securely and efficiently. (6 Create) – S413: Evaluates architectural options and justifies design choices with reference to technical, ethical, and regulatory requirements. (5 Evaluate) – S414: Collaborates with engineers, data scientists, and domain experts to design systems that are auditable and sustainable. (5 Evaluate) – S415: Implements design review and risk assessment processes to identify and mitigate potential harms. (5 Evaluate) Behaviour Statements: – B246: Demonstrates balanced judgement when making design trade-offs. (NA NA) – B247: Promotes responsible and sustainable design practices within teams. (NA NA) – B233: Demonstrates integrity, respect, and accountability in all aspects of professional conduct (2 Apply - V) – B236: Models ethical behaviour and encourages others to uphold professional standards (3 Apply - LA) C45: Effectively select, integrate, and contribute to open-source tools, frameworks, and libraries that support robust and reproducible AI system development. Proficiency Level: 6 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate
Consultation Draft for Review (December 2025). 49 Knowledge Statements: – K326: Understands the role and value of open-source software in accelerating AI innovation and transparency (2 Understand) – K327: Understands open-source licensing models, intellectual property rights, and attribution requirements (2 Understand) – K328: Understands dependency management, version control, and integration practices for open-source AI components (2 Understand) – K329: Understands criteria for evaluating the maturity, credibility, and sustainability of open-source projects (2 Understand) – K330: Understands principles of software quality, documentation, and testing in collaborative, open environments (2 Understand) Skill Statements: – S416: Identify, evaluate, and select appropriate open-source frameworks, libraries, and datasets for AI solutions (5 Evaluate) – S417: Integrate open-source components securely and effectively, ensuring compatibility and maintainability (6 Create) – S418: Contribute to open-source projects through code, documentation, testing, or review in line with project guidelines (6 Create) – S419: Use collaborative platforms (e.g., GitHub, GitLab) responsibly to manage contributions, issues, and version control. (3 Apply) – S420: Perform due diligence on open-source dependencies, assessing community activity, provenance, and support (5 Evaluate) Behaviour Statements: – B248: Adheres to open-source licensing, attribution, and contribution guidelines when selecting or integrating tools. (3 Apply - C) – B249: Communicates clearly when contributing to or modifying open-source components, ensuring changes are transparent and well-documented. (3 Apply - C) – B250: Collaborates constructively and respectfully with maintainers and contributors. (3 Apply - C) – B251: Demonstrates good stewardship by assessing the sustainability, security, and maintainability of open-source dependencies. (3 Apply - C) – B252: Promotes reproducibility by ensuring code, configuration, and dependencies are managed consistently. (3 Apply - C) C46: Engage with open-source communities in ways that uphold ethical, secure, and inclusive practices, promoting responsible participation in the AI ecosystem. Proficiency Level: 5 Lifecycle Stage(s): Build and Deploy | Operate and Evaluate Knowledge Statements: – K331: Understands ethical, social, and legal considerations in open-source collaboration, including privacy, consent, and responsible disclosure (2 Understand) – K332: Understands community governance structures, codes of conduct, and mechanisms for conflict resolution (2 Understand) – K333: Understands cybersecurity and supply-chain risks associated with open-source dependencies and contributions (2 Understand) – K334: Understands how diversity, accessibility, and equity influence the health and innovation of open-source communities (2 Understand) – K335: Understands responsible AI principles (e.g., fairness, accountability, transparency) as they apply to open collaboration. (2 Understand)
Consultation Draft for Review (December 2025). 50 Skill Statements: – S421: Apply ethical and secure development practices when using or contributing to open-source AI tools (3 Apply) – S422: Assess and mitigate risks related to security vulnerabilities, data misuse, and unethical code use (5 Evaluate) – S423: Engage respectfully and inclusively in global, cross-cultural open-source communities (3 Apply - C) – S424: Adhere to community codes of conduct, ensuring equitable participation and constructive communication (3 Apply - C) – S425: Promote responsible and transparent documentation of contributions to enhance auditability and trust (3 Apply - LA) Behaviour Statements: – B253: Engages respectfully and constructively with open-source communities, promoting an inclusive and supportive environment. (3 Apply - C) – B254: Raises security or ethical concerns responsibly, following community disclosure practices. (3 Apply) – B255: Acts in accordance with community codes of conduct, promoting fairness, accessibility, and respectful interaction. (3 Apply - C) – B256: Communicates transparently about contributions, decisions, and limitations to support trust and collaboration. (3 Apply - C) – B257: Demonstrates cultural awareness and sensitivity when engaging with global contributors. (3 Apply - C) C47: Collect and prepare the required datasets in line with organisation standards, ensuring accuracy, compliance and readiness for AI project use. Proficiency Level: 5 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K341: Understands the intended use and context of data in relation to the AI system’s purpose, stakeholders, and decision environment (2 Understand) – K342: Understands how data choices and preprocessing decisions influence model performance, interpretability, bias, and explainability (2 Understand) – K343: Understands the principles and techniques of data quality assessments, including completeness, consistency, validity, accuracy, and timeliness (2 Understand) – K344: Understands ethical and legal requirements related to data use, including privacy, consent, intellectual property, and data protection legislation (2 Understand) – K345: Understands governance frameworks for responsible data management, including transparency, auditability, and documentation standards (2 Understand) – K346: Understands risks of sampling bias, data imbalance, and representation gaps, and how these can affect fairness and equity of outcomes in AI systems (2 Understand) Skill Statements: – S431: Analyse the completeness and quality of available datasets (5 Evaluate) – S432: Construct datasets, potentially drawing from multiple disparate sources using data linkage (3 Apply) – S433: Perform data profiling and characterisation to understand the surface properties of the data (4 Analyse) – S434: Handle missing data, through principled inclusion/exclusion criteria and imputation methods (3 Apply) – S435: Take a systematic approach to data curation and the application of data quality controls (3 Apply)
Consultation Draft for Review (December 2025). 51 – S436: Evaluate whether available data supports the intended problem definition, stakeholder needs, and ethical constraints (5 Evaluate) – S437: Identify and address limitations in data coverage, sampling bias, and representational risks (5 Evaluate) – S438: Perform data cleaning and transformation, including handling outliers, formatting inconsistencies, and data standardisation (3 Apply) – S439: Prepare data in accordance with relevant governance frameworks to ensure traceability, auditability, and responsible practice (3 Apply) Behaviour Statements: – B261: Acts with integrity in handling data, respecting privacy, consent, and data ownership (3 Apply - V) – B262: Demonstrates accountability for the ethical and legal implications of data use (3 Apply - LA) – B263: Maintains transparency in data preparation decisions and communicates limitations clearly (3 Apply - LA) – B264: Applies a reflective and critical mindset, questioning data assumptions and representativeness (4 Analyse) – B265: Works collaboratively with stakeholders to ensure data preparation aligns with intended outcomes and responsible AI principles (3 Apply - C) C48: Support the application of AI and related technologies in ways that uphold rights, promote trust and deliver value. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design Knowledge Statements: – K317: Understands human rights, equality, and diversity principles as they apply to digital and AI contexts (2 Understand) – K318: Knows strategies for fostering public trust and meaningful engagement in the responsible adoption of AI (2 Understand) – K319: Understands methods for aligning AI use with organisational missions and societal value creation (2 Understand) – K320: Knows that ethical tensions may arise in AI practice and understand the mechanisms to assess and balance benefit, risk, and proportionality (2 Understand) – K321: Understands stakeholder expectations and accountability requirements for responsible AI deployment (2 Understand) Skill Statements: – S407: Applies rights-based and human-centric design principles to AI implementation (3 Apply) – S408: Builds strategies that enhance trust, transparency, and public understanding of AI (6 Create) – S409: Engages effectively with communities, partners, and regulators to align AI use with shared values (3 Apply - C) – S410: Articulates the business, ethical, and/or societal case for AI adoption in clear, balanced terms (6 Create) – S411: Measures and communicates the value created by AI in ways that include social and ethical dimensions (5 Evaluate) Behaviour Statements: – B241: Acts as a responsible steward of technology, prioritising fairness, inclusivity, and wellbeing (3 Apply - LA) – B242: Builds trust through openness, honesty, and accountability (3 Apply - V)
Consultation Draft for Review (December 2025). 52 – B243: Promotes AI as a force for public good, balancing innovation with care (3 Apply - LA) – B244: Engages with AI's potential to serve the public interest, while critically assessing risks, limitations and societal impact (5 Evaluate) – B245: Demonstrates empathy, professionalism, and respect in all interactions related to AI adoption (3 Apply - V) C49: Collaborate effectively in multidisciplinary AI development teams. Proficiency Level: 3 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K265: Understands the value of interdisciplinary collaboration in AI development and deployment. (2 Understand) – K266: Knows strategies for communicating technical information clearly to varied audiences. (1 Remember) – K267: Understands the roles, expertise, and perspectives of different professionals involved in the AI lifecycle. (2 Understand) – K311: Understands best practice in stakeholder communication and engagement (3 Apply) Skill Statements: – S345: Communicates effectively across disciplinary and organisational boundaries. (3 Apply - C) – S346: Facilitates inclusive collaboration that values diverse expertise and lived experience. (3 Apply - C) – S347: Contributes constructively to peer review, feedback, and shared decision-making. (3 Apply - C) – S348: Helps translate between technical and non-technical perspectives to ensure mutual understanding. (3 Apply - C) Behaviour Statements: – B201: Demonstrates openness to feedback and alternative perspectives. (3 Apply - V) – B202: Builds trust and mutual respect within teams and with stakeholders. (3 Apply - C) – B233: Demonstrates integrity, respect, and accountability in all aspects of professional conduct (2 Apply - V) – B236: Models ethical behaviour and encourages others to uphold professional standards (3 Apply - LA) C50: Implement, customise, and optimise low-code and no-code solutions that address organisational challenges, improve workflows, and enhance efficiency. Proficiency Level: 5 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K185: Features, capabilities, and limitations of low-code and no-code platforms. (3 Apply) – K186: Principles of workflow design, process automation, and integration. (2 Understand) – K187: Common data structures, connectors, and APIs used in automation tools. (2 Understand) – K188: Knowledge of security, compliance, and governance considerations when deploying low-code/no-code solutions. (2 Understand)
Consultation Draft for Review (December 2025). 53 – K189: Methods for testing, validating, and documenting system configurations. (2 Understand) Skill Statements: – S264: Configure low-code/no-code tools to meet user and business requirements. (3 Apply) – S265: Adapt existing workflows and templates to optimise processes. (3 Apply) – S266: Integrate low-code/no-code solutions with other systems and data sources. (3 Apply) – S267: Troubleshoot and resolve issues with configured solutions. (5 Evaluate) – S268: Document and communicate solution design and updates to stakeholders. (3 Apply - C) Behaviour Statements: – B126: Demonstrates a problem-solving and solution-oriented mindset. (3 Apply) – B128: Balances speed of delivery with accuracy, security, and quality. (3 Apply) – B129: Shows adaptability and willingness to experiment with new tools and approaches. (3 Apply - L) – B130: Engages users and stakeholders to ensure solutions are practical and effective. (3 Apply - C) C51: Advocate for and ensure automation solutions are implemented responsibly, with fairness, transparency, and consideration for workforce wellbeing. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K179: Explain the ethical principles underpinning AI and automation, including fairness, accountability, and transparency. (2 Understand) – K181: Describe the employment, organisational, and wider social implications of automation and AI adoption. (2 Understand) – K182: Identify and interpret relevant legislation, standards, and best practices that support responsible and compliant AI use. (5 Evaluate) – K183: Recognise and evaluate strategies for balancing efficiency gains from automation with employee experience, wellbeing, and organisational culture. (5 Evaluate) – K184: Demonstrate understanding of inclusive design and accessibility considerations in the development and deployment of AI systems. (2 Understand) Skill Statements: – S258: Identify and assess potential ethical, legal, or workforce risks in automation. (5 Evaluate) – S259: Advocate for balanced solutions that prioritise fairness and transparency. (3 Apply - C) – S260: Engage diverse stakeholders (including workforce representatives) in decisionmaking. (3 Apply - C) – S262: Design communication and training plans that build workforce confidence and support responsible adoption. (6 Create) – S263: Integrate responsible innovation practices into automation projects. (3 Apply - LA) Behaviour Statements: – B122: Acts with integrity and responsibility, considering long-term impacts of automation. (3 Apply)
Consultation Draft for Review (December 2025). 54 – B123: Balances organisational goals with workforce wellbeing and inclusion. (3 Apply - LA) – B124: Promotes openness and transparency in automation decisions. (3 Apply - V) – B125: Demonstrates empathy, fairness, and respect for diverse perspectives. (3 Apply - V) C52: Develop and apply methods to ensure AI explainability and interpretability to ensure that AI systems are transparent, interpretable, and understandable to appropriate audiences. Proficiency Level: 6 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K275: Understands key approaches to explainable AI (XAI), including feature importance, counterfactuals, and model-agnostic techniques. (2 Understand) – K276: Knows how explainability requirements vary with risk, application domain, and regulatory context. (2 Understand) – K277: Understands cognitive and communication factors influencing how users interpret AI explanations. (2 Understand) – K310: Knows the professional, legal, and societal responsibilities associated with AI and data use (2 Understand) Skill Statements: – S357: Selects and applies appropriate interpretability methods for different model types and audiences. (5 Evaluate) – S358: Designs interfaces and visualisations that make AI reasoning understandable to end-users. (6 Create) – S359: Communicates limitations, uncertainty, and potential sources of error in AI decisions. (3 Apply - C) – S360: Evaluates and iterates explainability approaches based on user feedback and trust metrics. (5 Evaluate) Behaviour Statements: – B207: Promotes transparency as a foundation of trustworthy AI. (3 Apply - LA) – B208: Acts responsibly when communicating explanations, avoiding overconfidence or misleading claims. (3 Apply - V) – B236: Models ethical behaviour and encourages others to uphold professional standards (3 Apply - LA) C53: Design and operate AI systems with due consideration for environmental and social sustainability. Proficiency Level: 6 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K284: Understands environmental and social impacts of AI systems, including energy consumption, carbon footprint, and resource use. (2 Understand) – K285: Knows frameworks for sustainable and responsible technology design. (1 Remember) – K286: Understands trade-offs between computational performance, accuracy, and sustainability. (2 Understand) – K310: Knows the professional, legal, and societal responsibilities associated with AI and
Consultation Draft for Review (December 2025). 55 data use (2 Understand) Skill Statements: – S369: Assesses and reports environmental impact metrics associated with AI systems. (5 Evaluate) – S370: Selects architectures, infrastructure, and models to reduce energy and resource use. (5 Evaluate) – S371: Integrates sustainability goals into AI project planning and evaluation. (6 Create) – S372: Advocates for responsible innovation that considers societal benefit and environmental stewardship. (3 Apply - LA) Behaviour Statements: – B213: Demonstrates environmental and social responsibility in professional decisionmaking. (3 Apply - LA) – B214: Encourages others to consider sustainability as a core dimension of AI practice. (3 Apply - LA) – B236: Models ethical behaviour and encourages others to uphold professional standards (3 Apply - LA) C54: Apply AI governance and risk management principles to ensure compliance and accountability. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K258: Knows key AI governance frameworks, assurance standards, and regulatory instruments (e.g. ISO 42001, EU AI Act, UK AI Assurance Framework). (1 Remember) – K259: Understands methods for identifying, assessing, and mitigating risks across the AI lifecycle. (2 Understand) – K260: Knows the importance of documentation, traceability, and auditability to demonstrate compliance. (2 Understand) – K310: Knows the professional, legal, and societal responsibilities associated with AI and data use (2 Understand) – K261: Understands how data and model quality influence fairness, reliability, and stakeholder impact. (4 Analyse) Skill Statements: – S337: Applies structured governance frameworks to AI development and deployment. (3 Apply) – S338: Conducts impact and risk assessments, identifying technical, ethical, and organisational risks. (5 Evaluate) – S339: Develops and maintains documentation to support audit and regulatory compliance. (6 Create) – S340: Advises colleagues and stakeholders on governance best practice and regulatory readiness. (3 Apply - LA) Behaviour Statements: – B197: Upholds transparency and accountability in all governance processes. (3 Apply - LA) – B198: Champions compliance and ethical practice as integral to innovation. (3 Apply - LA) – B233: Demonstrates integrity, respect, and accountability in all aspects of professional conduct (2 Apply - V) – B236: Models ethical behaviour and encourages others to uphold professional standards
Consultation Draft for Review (December 2025). 56 (3 Apply - LA) C55: Support the development of AI systems to proactively manage potential workforce impacts. Proficiency Level: 5 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K180: Understand the potential social and economic impacts of AI automation on different roles, particularly for non-technical staff. (2 Understand) – K191: Awareness of ethical principles and professional standards relevant to AI development (e.g., fairness, transparency, accountability, human dignity). (1 Remember) – K202: Knowledge of change management principles and how technology adoption affects workplace culture and employee wellbeing. (2 Understand) – K211: Awareness of strategies for inclusive communication with staff from diverse, nontechnical backgrounds. (1 Remember) – K222: Understand the importance of designing AI systems that augment rather than unnecessarily replace human work where feasible. (2 Understand) Skill Statements: – S128: Engage with non-technical staff to understand their roles, responsibilities, and concerns when automation solutions are proposed. (4 Analyse) – S238: Translate technical concepts into accessible language to support transparent dialogue with stakeholders. (2 Understand) – S250: Apply ethical and human-centred design principles when scoping, developing, and deploying AI solutions. (3 Apply) – S261: Assess risks and unintended consequences of automation projects, including their impact on job roles. (5 Evaluate) – S272: Collaborate with colleagues across business functions to identify opportunities for retraining, redeployment, or upskilling of affected staff. (3 Apply - C) Behaviour Statements: – B20: Demonstrates empathy by actively considering the perspectives and concerns of staff who may be impacted by AI-driven change. (3 Apply - C) – B104: Acts responsibly, balancing organisational efficiency goals with fairness to employees. (3 Apply) – B116: Shows openness and honesty in communication, avoiding unnecessary jargon or obfuscation. (3 Apply - C) – B127: Values inclusivity, ensuring all affected staff feel part of discussions around new AI systems. (3 Apply - C) – B138: Maintains professionalism when discussing sensitive workforce impacts, showing respect for individuals’ contributions. (3 Apply - C) C56: Support the responsible, ethical and safe adoption of AI. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K143: Familiarity with ethical frameworks and principles such as fairness, accountability, and transparency. (2 Understand) – K154: Understands the potential safety risks associated with AI systems, including unintended behaviours, misuse, and emergent harms. (2 Understand) – K155: Understanding of the regulatory and legal environment including GDPR, AI
Consultation Draft for Review (December 2025). 57 regulation, intellectual property, liability and associated risks of non-compliance. (2 Understand) – K161: Awareness of societal impacts of AI such as bias, job displacement, environmental costs and privacy risks. (2 Understand) Skill Statements: – S205: Embeds ethical assessment into AI project design through impact assessments, audits, and bias testing. (6 Create) – S216: Coordinates governance and oversight structures with clear lines of accountability. (3 Apply - LA) – S228: Develops and executes responses to AI-related incidents including mitigation and communication. (6 Create) Behaviour Statements: – B81: Acts with integrity by raising ethical concerns where needed. (3 Apply - LA) – B92: Demonstrates empathy, considering human and societal consequences of AI systems. (3 Apply - C) – B101: Models transparency in explaining assumptions, limitations, and system behaviours. (3 Apply - LA) – B102: Challenges unsafe practices and assumptions by raising concerns, regardless of role or seniority. (3 Apply - LA) C57: Ensure responsible governance, security and stewardship of data across the AI lifecycle. Proficiency Level: 6 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K307: Understands principles of data ethics, data protection, security, and information governance (2 Understand) – K308: Knows data lifecycle processes, including collection, cleaning, integration, retention, and disposal (2 Understand) – K310: Knows the professional, legal, and societal responsibilities associated with AI and data use (2 Understand) Skill Statements: – S114: Understand data provenance processes, including how data is sourced, documented, transformed, and traced across systems to support transparency, accountability, and reproducibility. (2 Understand) – S115: Identify infrastructure requirements for data storage and analysis (5 Evaluate) – S116: Implement data protection policies and procedures (3 Apply) – S118: Ensure safe and secure management of sensitive data, models and infrastructures (3 Apply) – S119: Understand the necessary steps to ensure the safe, secure and robust management of sensitive data, models and infrastructures and an awareness of how these link to professional responsibilities. (3 Apply) – S120: Understand the actions that should be taken to respond to potential data loss in line with organisational, legal and regulatory procedures. (2 Understand) – S121: Understand FAIR Guiding Principles for scientific data management and stewardship into practices, where appropriate and practicable. (2 Understand) – S122: Understand the relationship between technical standards and regulation/governance and their benefits for interoperability and knowledge sharing. (2 Understand) – S123: Identify and manage the risks of erroneous and biased data. (5 Evaluate)
Consultation Draft for Review (December 2025). 64 project lifecycle. Proficiency Level: 5 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K190: Understanding of principles of effective professional communication (written, verbal, and visual). (2 Understand) – K192: Knowledge of approaches to apply management techniques and reporting frameworks. (2 Understand) – K193: Knowledge of tools and channels for collaborative communication and project reporting. (2 Understand) – K194: Understand the importance of transparency, accountability, and audience tailoring in communication. (2 Understand) – K195: Familiarity with organisational policies and expectations around reporting and governance. (2 Understand) Skill Statements: – S269: Prepare concise and clear progress updates tailored to different audiences. (3 Apply - C) – S270: Use appropriate communication methods (presentations, dashboards, written reports, meetings). (3 Apply - C) – S271: Identify and escalate risks or issues in a timely manner. (5 Evaluate) – S273: Engage stakeholders through active listening and constructive dialogue. (3 Apply - C) – S274: Document decisions, actions, and outcomes for accountability and future reference. (3 Apply - LA) Behaviour Statements: – B131: Demonstrates openness, honesty, and professionalism in communication. (3 Apply - V) – B132: Keeps stakeholders engaged, informed, and confident in progress being made. (3 Apply - C) – B133: Adapts communication style to suit diverse audiences. (3 Apply - C) – B134: Takes responsibility for ensuring stakeholders are informed, with particular attention to those most impacted. (3 Apply - C) C67: Design and develop data-driven and AI solutions that are intuitive and human-centred. Proficiency Level: 6 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K9: Understands human-centred design principles and their application to data-driven and AI solutions. (2 Understand) – K10: Knows the importance of involving users and stakeholders throughout the design and development process to ensure relevance and usability. (2 Understand) – K11: Understands inclusive and accessible design principles to create AI systems that serve diverse users and contexts equitably. (2 Understand) – K13: Knows the cognitive, social, and emotional factors that influence human interaction with AI systems. (2 Understand) – K14: Understands the role of transparency, explainability, and trust in encouraging user confidence in AI solutions. (2 Understand) – K15: Knows how to apply human–AI teaming and human-in-the-loop design approaches
Consultation Draft for Review (December 2025). 65 to balance automation with human judgement. (2 Understand) – K16: Understands usability testing methodologies and how to translate findings into iterative design improvements. (2 Understand) – K17: Knows the ethical and societal implications of AI design decisions, including fairness, accountability, and unintended consequences. (2 Understand) – K18: Understands how design decisions influence system adoption, effectiveness, and long-term sustainability. (2 Understand) Skill Statements: – S2: Implement human-factors considerations with respect to data-driven and AI-powered solutions. (3 Apply) – S11: Evaluate appropriate approaches to augment human skills with artificial intelligence (e.g. ‘human-in-the-loop’ and ‘human–AI teaming’). (5 Evaluate) – S12: Apply user-centred design principles to guide solution development. (3 Apply) – S13: Design intuitive and user-friendly interfaces to facilitate interaction between endusers and AI systems. (6 Create) – S14: Perform usability testing of AI system interfaces and interpret results to improve user experience. (4 Analyse) – S15: Ensure that data-driven and AI systems are used in an ethical and responsible way. (5 Evaluate) – S16: Engage diverse stakeholders, including end-users, to co-design and validate AI solutions that meet real-world needs. (3 Apply - C) – S17: Apply inclusive design principles to ensure accessibility and equity in AI-enabled systems. (3 Apply - C) – S18: Translate user research and behavioural insights into system requirements and interface design decisions. (4 Analyse) – S3: Evaluate cognitive load, interpretability, and trust in AI–human interactions to optimise usability and understanding. (5 Evaluate) – S4: Apply responsible innovation practices to anticipate and mitigate potential societal or ethical impacts of AI systems. (5 Evaluate) – S5: Integrate explainability and feedback mechanisms into system design to support user confidence and oversight. (6 Create) Behaviour Statements: – B13: Champion user-centred and ethical design principles throughout the development of AI solutions. (3 Apply - LA) – B14: Seek diverse perspectives to ensure inclusivity and accessibility in AI design. (3 Apply - C) – B15: Demonstrate empathy and curiosity in understanding user needs and translating them into effective design choices. (3 Apply - C) – B2: Take accountability for the societal and human impacts of AI design decisions. (3 Apply - C) – B3: Embrace iterative design and feedback to continuously improve usability and user trust. (3 Apply - L) C68: Collaborate across disciplines to identify risks, integrate diverse perspectives and support responsible innovation. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K295: Understands interdisciplinary approaches to AI, including social science, legal, policy, and technical perspectives (2 Understand)
Consultation Draft for Review (December 2025). 66 – K296: Knows frameworks for responsible innovation and stakeholder co-creation (2 Understand) – K297: Understands group dynamics, communication techniques, and inclusive practices in cross-functional collaboration (2 Understand) – K298: Knows the types of risk associated with AI adoption, including ethical, reputational, safety, and societal risks (2 Understand) Skill Statements: – S386: Leads or contributes to multidisciplinary assessments of emerging risks and opportunities (3 Apply - C) – S382: Facilitates constructive collaboration across diverse professional and disciplinary boundaries (3 Apply - C) – S383: Translates technical concepts into accessible language for non-specialist audiences (3 Apply - C) – S384: Identifies and integrates ethical, legal, and social implications into decision-making (6 Create) – S385: Builds consensus and alignment among stakeholders with differing priorities (3 Apply - C) Behaviour Statements: – B221: Values diversity of expertise, thought, and lived experience (3 Apply - V) – B222: Demonstrates openness, curiosity, and respect for differing viewpoints (3 Apply - V) – B223: Acts as a connector between technical teams, business leaders, and wider communities (3 Apply - C) – B224: Promotes an environment that encourages challenge, reflection, and continuous learning (3 Apply - L) C69: Research, critically evaluate, and apply emerging knowledge to AI automation developments, ensuring practice remains current, ethical, and forward-looking. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K196: Identify emerging trends, tools, and technologies in AI and automation. (2 Understand) – K197: Describe industry standards, regulations, and best practices relevant to AI adoption. (2 Understand) – K198: Understand the ethical, legal, and societal implications of AI automation. (2 Understand) – K199: Outline methods for horizon scanning, professional networking, and continuous learning. (2 Understand) – K200: Understanding of risk assessment frameworks and approaches for evaluating new technologies. (2 Understand) Skill Statements: – S275: Research and critically appraise new developments in AI automation. (5 Evaluate) – S276: Translate insights on trends and risks into actionable recommendations for practice. (6 Create) – S277: Evaluates opportunities for innovation and process improvement where AI automation is appropriate. (5 Evaluate) – S278: Engage with professional communities and knowledge sources to maintain
Consultation Draft for Review (December 2025). 67 expertise. (3 Apply - C) – S279: Apply risk assessment methods to anticipate and mitigate potential issues. (5 Evaluate) Behaviour Statements: – B135: Demonstrates curiosity, openness, and a proactive approach to continuous learning. (3 Apply - L) – B136: Maintains a responsible and ethical mindset, remaining unbiased in when considering and evaluating new opportunities. (3 Apply - V) – B137: Shares knowledge with colleagues and stakeholders to build collective capability. (3 Apply - L) – B139: Balances innovation with caution, considering both benefits and risks. (5 Evaluate) C70: Evaluate emerging AI technologies and their potential societal impact. Proficiency Level: 5 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K268: Understands current and emerging trends in AI research, technology, and application. (2 Understand) – K269: Knows frameworks for assessing societal and ethical implications of new technologies. (1 Remember) – K270: Understands principles of responsible innovation and anticipatory governance. (2 Understand) – K310: Knows the professional, legal, and societal responsibilities associated with AI and data use (2 Understand) Skill Statements: – S349: Scans and analyses emerging AI technologies to assess readiness, risks, and opportunities. (5 Evaluate) – S350: Advises organisations on responsible adoption and experimentation. (3 Apply - LA) – S351: Communicates potential impacts clearly to support evidence-based decisionmaking. (3 Apply - C) – S352: Contributes to foresight, research, or policy dialogues on the future of AI. (3 Apply - LA) Behaviour Statements: – B203: Demonstrates curiosity and openness to new ideas and technologies. (3 Apply - V) – B204: Acts as a responsible innovator, balancing progress with precaution. (3 Apply - V) – B233: Demonstrates integrity, respect, and accountability in all aspects of professional conduct (2 Apply - V) – B236: Models ethical behaviour and encourages others to uphold professional standards (3 Apply - LA) C71: Support learning, mentorship, and capacity building in responsible and ethical AI practice. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K287: Understands principles of effective mentorship, facilitation, and learning in
Consultation Draft for Review (December 2025). 68 professional settings. (2 Understand) – K288: Knows frameworks for AI literacy and capability development. (1 Remember) – K289: Understands how to communicate technical and ethical concepts to varied audiences. (2 Understand) – K311: Understands best practice in stakeholder communication and engagement (3 Apply) Skill Statements: – S373: Provides mentorship and constructive feedback to peers, leadership and junior colleagues. (3 Apply - C) – S374: Designs or contributes to training and awareness activities on responsible AI. (6 Create) – S375: Shares knowledge and best practice within professional networks and communities. (3 Apply - C) – S376: Promotes continuous learning and reflection on emerging technologies and ethical challenges. (3 Apply - LA) Behaviour Statements: – B215: Demonstrates generosity, humility, and curiosity in sharing knowledge. (3 Apply - V) – B216: Encourages others to grow in confidence and capability. (3 Apply - C) – B236: Models ethical behaviour and encourages others to uphold professional standards (3 Apply - LA) C72: Engage with and/or develop governance processes that enable validation, oversight and escalation. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K299: Understands principles of corporate governance, risk management, and assurance as they relate to AI (2 Understand) – K300: Knows governance frameworks such as model risk management, impact assessment, and ethical review (2 Understand) – K301: Understands escalation pathways, accountability structures, and decision rights for AI oversight (2 Understand) – K302: Knows internal and external audit requirements for AI and data-driven systems (2 Understand) Skill Statements: – S387: Designs, implements, or contributes to governance frameworks that ensure proportionate oversight of AI systems (6 Create) – S388: Defines and applies performance indicators and control mechanisms for monitoring AI use (6 Create) – S389: Escalates risks, anomalies, or ethical concerns through established governance channels (3 Apply - LA) – S390: Provides evidence and documentation to support audit, assurance, and compliance activities (3 Apply - C) – S391: Reviews and refines governance processes to adapt to changing technologies or policies (5 Evaluate) Behaviour Statements: – B225: Acts transparently and decisively when governance concerns arise (3 Apply - LA) – B226: Promotes accountability, fairness, and consistency in oversight processes (2
Consultation Draft for Review (December 2025). 69 Apply - LA) – B227: Encourages open reporting and continuous improvement in governance practice (3 Apply - LA) – B228: Demonstrates professional courage in challenging poor practice or undue influence (3 Apply - V) C73: Evaluate and oversee external AI vendors and third-party tools to ensure technical fitness, compliance, and responsible operation. Proficiency Level: 5 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K336: Understands common third-party AI solution architectures, including APIs, proprietary models, SaaS platforms, and embedded AI features. (2 Understand) – K337: Knows the technical and data governance risks associated with external AI systems (e.g., data sovereignty, model opacity, update cycles, dependency risks). (2 Understand) – K338: Awareness of organisational procurement standards and assurance requirements relevant to AI tools. (2 Understand) – K339: Understands the evidential requirements for validating vendor claims about model performance, fairness, robustness, and security. (2 Understand) – K340: Awareness of the legal and regulatory context affecting third-party AI tools, including GDPR, intellectual property, and AI-specific regulatory obligations. (1 Remember) Skill Statements: – S426: Conducts technical assessments of vendor tools, including validation of performance, accuracy, bias, and safety claims. (5 Evaluate) – S427: Identifies integration risks, data requirements, and operational constraints when adopting third-party AI systems. (4 Analyse) – S428: Reviews technical documentation, audit artefacts, and assurance materials provided by vendors. (5 Evaluate) – S429: Raises technical, ethical, or data concerns to procurement, compliance, or leadership teams where vendor solutions present risks. (3 Apply - LA) – S430: Monitors live system behaviour, reporting anomalies, performance drift, or noncompliant vendor updates. (4 Analyse) Behaviour Statements: – B258: Demonstrates diligence and objectivity when reviewing vendor claims, critically evaluating marketing material and vendor claims. (3 Apply - V) – B259: Acts responsibly when identifying risks, ensuring concerns are escalated in a timely and evidence-based manner. (3 Apply - LA) – B260: Values transparency, documenting findings and sharing knowledge across teams. (3 Apply - C) C74: Define and communicate a coherent AI strategy that aligns with organisational goals and values. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design Knowledge Statements: – K217: Understands the organisation’s vision, mission, values, and long-term objectives, and can identify if and how AI can support and accelerate their achievement. (2
Consultation Draft for Review (December 2025). 70 Understand) – K218: Has a strong awareness of the wider AI landscape, including current technological capabilities, limitations, sector-specific technologies, and risks. (2 Understand) – K219: Understands the ethical, legal, and regulatory requirements that shape responsible AI adoption, including principles of fairness, accountability, and transparency. (2 Understand) – K220: Knows established frameworks and methodologies for developing and aligning strategy, including how to balance innovation with risk management, compliance, and organisational values. (2 Understand) – K221: Understands approaches to change management and stakeholder communication that are effective in gaining buy-in for AI adoption. (2 Understand) – K223: Is knowledgeable about financial and resource planning processes that ensure the AI strategy is realistic, sustainable, and scalable. (2 Understand) Skill Statements: – S296: Can translate high-level organisational priorities into a coherent AI strategy, with clear goals, measurable outcomes, and success indicators. (6 Create) – S297: Demonstrates the ability to identify, evaluate, and prioritise AI opportunities, balancing ambition with technical feasibility, ethical responsibility, and organisational readiness. (5 Evaluate) – S298: Can build consensus across senior leadership, presenting the case for AI adoption in ways that resonate with both technical and non-technical audiences. (3 Apply - C) – S299: Is able to communicate complex AI concepts clearly, tailoring explanations to different stakeholders across the organisation. (2 Understand) – S300: Can establish governance structures, operating models, and processes that embed the AI strategy into day-to-day operations. (6 Create) – S301: Demonstrates the ability to benchmark pace of development and capability against peers, competitors, and industry standards, and adapt the strategy accordingly. (4 Analyse) Behaviour Statements: – B152: Acts consistently as a role model for ethical, responsible, and values-driven AI adoption. (3 Apply - LA) – B153: Demonstrates strategic vision and long-term thinking while remaining agile and responsive to emerging risks and opportunities. (3 Apply - LA) – B154: Builds trust and credibility by engaging openly with stakeholders, listening to concerns, and addressing risks transparently. (3 Apply - LA) – B155: Encourages a culture of innovation and continuous improvement, while keeping focus on delivering tangible organisational value. (3 Apply - LA) – B156: Shows resilience and accountability in leading AI strategy through uncertainty, ambiguity, and resistance to change. (3 Apply - LA) C75: Balance innovation with practical delivery in AI adoption. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K224: Understands how innovation methodologies such as iterative development, piloting, and scaling approaches apply to AI, and recognises how to adapt them within the constraints of organisational structures and resources. (2 Understand) – K225: Is aware of the organisational realities that shape AI adoption, including financial and budgetary constraints, regulatory obligations, workforce readiness, and cultural attitudes towards risk and change. (2 Understand) – K226: Has knowledge of technology readiness levels (TRLs) and innovation pipeline
Consultation Draft for Review (December 2025). 71 maturity, and can use these frameworks to assess when it is appropriate to transition AI initiatives from experimentation into production. (2 Understand) – K227: Understands the trade-offs between encouraging ambitious, cutting-edge AI innovation and ensuring that delivery remains feasible, measurable, and aligned to business priorities. (2 Understand) Skill Statements: – S302: Can design and oversee AI adoption programmes that deliberately balance exploratory innovation with practical delivery, ensuring that neither creativity nor operational impact is compromised. (6 Create) – S304: Demonstrates the ability to plan and manage pilots, proofs of concept, and phased roll-outs, progressively validating AI solutions before committing to full-scale deployment. (6 Create) – S305: Is able to evaluate competing demands between rapid innovation and operational stability, and make reasoned judgements about where to allocate resources and focus. (5 Evaluate) – S306: Can create and manage pathways for successful innovations to transition into core operations, ensuring lessons learned in experimentation are applied to long-term delivery. (6 Create) Behaviour Statements: – B157: Demonstrates pragmatism and adaptability when innovation initiatives need to be adjusted to fit organisational realities or delivery constraints. (3 Apply - V) – B158: Champions innovation within the organisation while taking accountability for ensuring that ideas translate into reliable and valuable outcomes. (3 Apply - LA) – B159: Encourages experimentation and creativity within teams, while consistently linking innovation back to strategic organisational objectives to maintain focus and discipline. (3 Apply - LA) C76: Sponsor governance frameworks that safeguard lawful, ethical and responsible AI use. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K228: Has a comprehensive understanding of the legal and regulatory frameworks that govern AI use, including data protection laws, equality and non-discrimination legislation, industry-specific regulations, and emerging standards for trustworthy AI. (2 Understand) – K229: Understands the principles of corporate governance and how they apply specifically to AI systems, including the creation of accountability structures, clear lines of decision-making, and oversight mechanisms. (2 Understand) – K230: Is knowledgeable about best practices in AI assurance, such as testing for fairness, quality, mitigating bias, implementing robust monitoring systems, and ensuring transparency of AI operations. (2 Understand) – K231: Appreciates how governance frameworks must evolve over time to remain relevant as technology, regulation, and societal expectations change. (2 Understand) Skill Statements: – S307: Can design AI governance frameworks that integrate legal, ethical, and operational safeguards, ensuring compliance while also supporting innovation. (6 Create) – S308: Demonstrates the ability to embed governance processes into existing enterprise risk management and compliance functions, aligning AI oversight with broader organisational systems rather than treating it as an isolated activity. (6 Create)
Consultation Draft for Review (December 2025). 72 – S309: Is able to establish reporting structures, documentation standards, and audit processes that provide clear accountability and traceability for AI systems throughout their lifecycle. (6 Create) – S310: Can evaluate the adequacy of existing governance structures and adapt them responsively to address new risks, regulations, or technological developments. (5 Evaluate) Behaviour Statements: – B161: Acts with integrity and fairness when establishing or applying governance measures, ensuring that decision-making processes are transparent and defensible. (3 Apply - V) – B162: Consistently promotes the importance of compliance and responsible use as shared organisational responsibilities, not simply the concern of technical teams or product owners. (3 Apply - LA) – B163: Demonstrates openness and consistency in implementing governance standards, encouraging trust both internally and externally in the organisation’s use of AI. (3 Apply - LA) C77: Embed ethical principles in AI decision-making and operations. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K232: Understands the range of ethical frameworks that apply to AI, including fairness, accountability, transparency, inclusivity, sustainability, and respect for human autonomy. (2 Understand) – K234: Knows how ethical considerations can be embedded across the entire AI lifecycle, from the initial design and procurement stage, through model development and deployment, to monitoring and continuous improvement. (2 Understand) – K235: Is aware of the potential tensions and trade-offs between ethical principles and short-term organisational or commercial goals, and recognises the importance of managing these tensions transparently. (2 Understand) Skill Statements: – S311: Demonstrates the ability to translate high-level ethical guidelines into operational standards, policies, and practices that are actionable and measurable within an organisation. (6 Create) – S312: Can facilitate cross-disciplinary dialogue, ensuring that ethical perspectives are considered not only by technical specialists as well as by business leaders, legal advisors, and affected communities. (3 Apply - C) – S313: Is able to monitor and audit AI systems to ensure ongoing adherence to ethical standards, and can initiate remedial action where systems fall short. (5 Evaluate) – S315: Can champion processes that enable ethical reflection and deliberation in day-today decision-making, helping teams to navigate through uncertainty and competing priorities responsibly. (3 Apply - LA) Behaviour Statements: – B164: Acts consistently as a values-driven leader, modelling ethical behaviour in their own decision-making and setting expectations for others to follow. (3 Apply - LA) – B165: Encourages openness and transparency in discussing ethical challenges, recognising that surfacing potential issues early is preferable to concealing them. (3 Apply - LA) – B166: Prioritises long-term trust, organisational reputation, and societal benefit, ensuring these are not compromised by short-term convenience or commercial gain (3 Apply - V)
Consultation Draft for Review (December 2025). 73 C78: Select AI vendors and suppliers through fair and evidence-based evaluation. Proficiency Level: 6 Lifecycle Stage(s): Design | Build and Deploy Knowledge Statements: – K236: Has a strong understanding of organisational procurement frameworks, including legal and regulatory requirements for competitive tendering, due diligence, and evidence-based evaluation. (2 Understand) – K237: Understands the risks associated with AI vendor and supplier relationships, including issues such as vendor lock-in, dependency on proprietary systems, bias in supplied models or datasets, ethical standards of third parties, and vulnerabilities in supply chains. (2 Understand) – K238: Is knowledgeable about the criteria for evaluating AI products and services, including technical quality, interoperability, scalability, security, compliance with standards, and the vendor’s commitment to ethical and responsible practices. (2 Understand) – K239: Recognises the importance of transparency and accountability in procurement processes, both for internal governance and to maintain public and stakeholder trust. (2 Understand) Skill Statements: – S316: Can design and apply structured, fair evaluation frameworks that enable objective comparisons between AI vendors and suppliers, based on both technical and nontechnical criteria. (6 Create) – S317: Demonstrates the ability to critically assess evidence provided by vendors, including claims about performance, accuracy, explainability, or compliance, and to validate those claims through independent testing or external benchmarks. (5 Evaluate) – S318: Is able to manage the procurement process in a way that ensures transparency, decision making traceability, and fairness, including the documentation of decisions and clear communication of rationale to stakeholders. (3 Apply - LA) – S319: Can negotiate effectively with vendors and suppliers to secure value for the organisation while maintaining fairness, impartiality, and alignment with organisational values. (3 Apply - LA) Behaviour Statements: – B167: Upholds impartiality and fairness in all procurement decisions, actively guarding against bias, favouritism, or undue influence. (3 Apply - V) – B168: Challenges conflicts of interest, inappropriate lobbying, or attempts to compromise the integrity of supplier selection. (3 Apply - V) – B169: Consistently aligns supplier selection processes with the organisation’s values, ethics, and long-term strategic objectives. (5 Evaluate) – B170: Models accountability and transparency, ensuring that procurement decisions can stand up to both internal and external scrutiny. (5 Evaluate) C79: Select technology platforms that meet organisational needs and mitigate long-term risks. Proficiency Level: 5 Lifecycle Stage(s): Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K240: Understands enterprise architecture principles, including interoperability, integration requirements, and the importance of aligning technology platforms with
Consultation Draft for Review (December 2025). 80 stakeholders, from technical experts to senior policymakers. (3 Apply - C) Behaviour Statements: – B30: Acts with professionalism, diplomacy, and credibility in external engagements, ensuring that the organisation is seen as a trusted and authoritative voice. (3 Apply - V) – B31: Advocates in ways that are consistent with organisational values and societal benefit, resisting the temptation to promote narrow or short-term interests. (3 Apply - LA) – B32: Builds long-term, trust-based relationships with policymakers and regulators, based on openness, integrity, and reliability. (3 Apply - C) – B33: Demonstrates humility by recognising the organisation’s role within a broader ecosystem of stakeholders and engaging collaboratively rather than competitively. (3 Apply - V) C87: Build trust with the public and external communities on AI use. Proficiency Level: 3 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K80: Understands societal concerns, debates, and expectations surrounding AI, including issues such as fairness, accountability, privacy, transparency, and the future of work. (2 Understand) – K82: Has knowledge of effective methods for engaging the public and external communities in dialogue about AI, such as consultation processes, participatory design, and open communication campaigns. (2 Understand) – K83: Recognises how organisational reputation and legitimacy are shaped by public trust, and how trust can be strengthened or undermined by the way the topic of AI is communicated and governed. (2 Understand) – K84: Is aware of the importance of tailoring messages to different communities, ensuring that engagement is meaningful and accessible. (3 Apply - C) Skill Statements: – S165: Can engage external stakeholders and communities in constructive, two-way dialogue about the organisation’s use of AI, actively listening as well as informing. (3 Apply - C) – S166: Demonstrates the ability to communicate complex technical topics in humancentred, accessible ways that resonate with diverse audiences. (3 Apply - C) – S167: Is able to demonstrate accountability by explaining the benefits of AI as well as its risks and limitations, along with showing how the organisation addresses them. (3 Apply - LA) – S168: Can build and sustain relationships with community groups, advocacy organisations, and other external stakeholders, working to position the organisation as a more transparent, accountable, and responsible actor. (3 Apply - C) Behaviour Statements: – B34: Acts with openness, honesty, and empathy when engaging with the public, ensuring that communication is not defensive or dismissive but genuinely inclusive. (3 Apply - LA) – B35: Demonstrates respect for diverse perspectives, recognising that communities may have different experiences, values, and levels of trust in technology. (3 Apply - V) – B36: Acts with integrity and consistency in ethical, responsible, and socially beneficial AI use, contributing to long-term trust. (3 Apply - V) – B37: Models humility by acknowledging uncertainties and limitations, showing that the organisation is willing to learn and adapt. (3 Apply - V)
Consultation Draft for Review (December 2025). 81 C88: Monitor regulatory changes and prepare the organisation for compliance. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K85: Has a comprehensive understanding of current AI regulations, industry standards, and guidance at national and international levels, including data protection, safety, liability, and sector-specific rules. (2 Understand) – K86: Understands how the global regulatory landscape for AI is evolving, with awareness of trends such as the EU AI Act, NIST AI Risk Management Framework, OECD principles, and national frameworks. (2 Understand) – K87: Knows how to conduct horizon scanning and regulatory foresight, identifying emerging legal, ethical, and policy issues that may affect organisational AI use. (4 Analyse) – K88: Recognises the organisational implications of regulatory change, including the need for adaptation of policies, processes, and technical systems. (4 Analyse) Skill Statements: – S169: Can systematically monitor regulatory developments and assess their potential impact on the organisation’s operations, strategy, and AI systems. (5 Evaluate) – S170: Demonstrates the ability to develop proactive compliance strategies that prepare the organisation in advance of regulatory deadlines, avoiding last-minute or reactive adjustments. (6 Create) – S171: Is able to advise senior leadership on regulatory risks and opportunities, enabling informed decisions about investment, risk appetite, and organisational positioning. (3 Apply - C) – S172: Can build organisational capacity to remain agile in the face of regulatory uncertainty, ensuring readiness for multiple possible futures. (3 Apply - LA) Behaviour Statements: – B38: Demonstrates agility and strategic awareness, treating regulatory readiness as an enabler rather than a constraint. (3 Apply - LA) – B39: Maintains accountability for ensuring that the organisation is prepared for change, modelling responsibility in the face of uncertainty. (3 Apply - LA) – B40: Encourages proactive engagement with regulators, industry bodies, and peers to share knowledge, influence debates, and anticipate shifts. (3 Apply - LA) C89: Oversee the embedding of regulatory readiness into business processes. Proficiency Level: 6 Lifecycle Stage(s): Framing and Feasibility | Design | Build and Deploy | Operate and Evaluate Knowledge Statements: – K89: Understands compliance frameworks, audit requirements, and risk management standards, and how they can be operationalised within everyday organisational processes. (2 Understand) – K90: Has knowledge of sector-specific regulatory obligations (e.g., healthcare, finance, energy) and of cross-border complexities where AI solutions operate internationally. (2 Understand) – K91: Recognises how business processes must be designed to enable compliance by default, reducing the reliance on manual interventions or ad hoc fixes. (4 Analyse) – K93: Understands the relationship between regulatory readiness, organisational resilience, and stakeholder trust. (2 Understand)
Consultation Draft for Review (December 2025). 82 Skill Statements: – S173: Can integrate regulatory compliance requirements into core business processes, ensuring that adherence is embedded rather than treated as an afterthought. (6 Create) – S174: Demonstrates the ability to design workflows, documentation standards, and system-level controls that provide evidence of compliance in a clear and auditable manner. (6 Create) – S176: Is able to train and support staff across the organisation so that compliance responsibilities are understood, shared, and consistently implemented. (3 Apply - LA) – S177: Can establish monitoring and feedback mechanisms that ensure compliance processes remain effective as regulations evolve. (6 Create) Behaviour Statements: – B43: Prioritises consistency and rigour in embedding compliance, demonstrating that it is a continuous responsibility rather than an occasional exercise. (3 Apply - V) – B44: Acts with attention to detail and a strong sense of responsibility, recognising that small lapses can create significant organisational risks. (3 Apply - LA) – B45: Promotes a culture in which compliance is seen as a shared value, contributing to trustworthiness and resilience, rather than as a bureaucratic burden. (3 Apply - LA) – B46: Models openness and transparency in demonstrating compliance to internal and external stakeholders. (3 Apply - LA) C90: Supports peers and communities to understand and use AI responsibly through clear guidance and examples. Proficiency Level: 3 Lifecycle Stage(s): Operate and Evaluate Knowledge Statements: – K159: Understands that not everyone has equal access to knowledge, confidence, or resources to engage safely with AI-enabled technologies. (2 Understand) – K160: Knows how to explain AI concepts, risks, and benefits in plain, accessible language suited to different audiences (e.g. children, older adults, or community groups). (2 Understand) Skill Statements: – S237: Can provide clear, practical demonstrations of AI tools (e.g. chatbots, recommendation systems, image generators) while highlighting both benefits and limitations. (2 Understand) – S240: Can help others identify trustworthy sources of information about AI, including official guidance, educational resources, and fact-checking tools. (3 Apply - L) Behaviour Statements: – B100: Acts as a trusted peer mentor, creating a supportive environment for questions and discussion without judgement. (3 Apply - LA) C91: Shares AI knowledge and practices in ways that are understandable to colleagues of different backgrounds. Proficiency Level: 3 Lifecycle Stage(s): Operate and Evaluate Knowledge Statements: – K157: Recognises that peer-to-peer learning and informal support can play an important role in building AI literacy across families, workplaces, and communities. (3 Apply - L)
Consultation Draft for Review (December 2025). 83 – K158: Understands that teaching others about AI requires empathy, patience, and respect for diverse experiences, including concerns about technology or privacy. (2 Understand) Skill Statements: – S236: Can adapt explanations to the audience’s needs, using examples from everyday life to make AI concepts relatable and relevant. (3 Apply - L) Behaviour Statements: – B98: Demonstrates patience, openness, and encouragement when helping others learn about AI. (3 Apply - L) – B99: Demonstrates a commitment to ongoing development to increase personal knowledge and experience, while supporting others to do the same. (3 Apply - L) C92: Helps others understand their data and consumer rights and encourages the use of service settings that prevent unnecessary sharing of user data or contribution of content to model training. Proficiency Level: 3 Lifecycle Stage(s): Operate and Evaluate Knowledge Statements: – K19: Understands the basics of data protection, privacy, and consumer rights, including the right to explanation and to challenge decisions made using AI. (2 Understand) – K20: Knows where individuals can seek trustworthy advice or redress regarding data use, privacy breaches, or AI-related harms. (2 Understand) – K21: Understands how AI systems collect and process data, and how this relates to transparency, consent, and accountability. (2 Understand) – K22: Knows that default settings (such as those controlling data sharing, retention, and contribution to AI model training) can significantly influence safety, accessibility, and user trust in digital and AI-enabled services. (2 Understand) Skill Statements: – S6: Can point people to trustworthy help (e.g. official guidance, helplines) and explain options in simple, accessible terms. (3 Apply - C) – S7: Can help others recognise when their rights might have been affected and identify appropriate next steps. (3 Apply - C) – S8: Can interpret and explain privacy or terms-of-service information in a clear and balanced way. (3 Apply - C) – S9: Can provide constructive feedback or raise concerns with service providers to advocate for safer, clearer defaults, including settings that minimise unnecessary data sharing or opt-in use of personal content for AI training. (3 Apply - C) – S10: Can communicate complex ideas about rights and data use in ways that are inclusive and sensitive to differing levels of digital confidence. (3 Apply - C) Behaviour Statements: – B4: Advocates for safer defaults (such as privacy-protective and non-training-datasharing settings) and clearer communication when helping others use AI-enabled services. (3 Apply - LA) – B5: Promotes fairness, respect, and inclusion when supporting others to understand their rights. (3 Apply - V) – B6: Acts with integrity and care when sharing advice or signposting to support. (3 Apply - V) – B7: Encourages organisations and service providers to adopt transparent and responsible design practices, including minimising default data sharing and clearly
Consultation Draft for Review (December 2025). 84 explaining when user content may be used to train AI models. (3 Apply - LA)
Consultation Draft for Review (December 2025). 85 Annex A: AI Project Lifecycle Stages One of the major developments in Version 3 of the AI Skills for Business Competency Framework is the introduction of an embedded AI Project Lifecycle. This addition is designed to help organisations, policymakers and practitioners understand how skills requirements develop as an AI system progresses from strategic framing to operational deployment. By aligning competencies to these stages, the framework clarifies how skills shift over time, who must contribute at each point, and how organisations can steward AI systems responsibly from concept through to retirement. In establishing the AI Project Lifecycle Stages for the AI Skills for Business Competency Framework, we reviewed a range of established lifecycle models for AI and data-driven systems, including those presented in ISO/IEC 22989, ISO/IEC 23053, ISO/IEC 23895, and the OECD Framework for the Classification of AI Systems. By synthesising rather than adopting an existing lifecycle model, we ensure that the lifecycle structure serves the central purpose of this framework: to provide a coherent, skills-aligned view of how data-driven and AI systems are developed, in terms which are accessible to the full breadth of beneficiaries of the AI Skills for Business Competency Framework. The lifecycle is therefore not a competing standard, but a practical bridge between established guidance and the competencies required to implement it responsibly. Full mapping to the aforementioned standards will be made available with the post-consultation version of the Framework. Here we present the AI Project Lifecycle, demonstrating the various phases required to move a data-driven solution and/or AI system from the ideation stage to full operational deployment. Our Lifecycle comprises four lifecycle stages; Framing and Feasibility, Design, Build and Deploy, and Operate and Evaluate. The AI Project Lifecycle is iterative, and a project is reasonably expected to move non-linearly between these stages. Our framework is depicted in the figure below, and each lifecycle stage is described in further detail in the remainder of this section.
Consultation Draft for Review (December 2025). 86 Problem Definition This phase focuses on exploring if data-driven and AI approaches have the potential to deliver effective solutions to support business objectives and organisations’ strategic goals. This phase entails identifying the business challenge and clearly defining the problem the organisation is trying to solve, a long with understanding why it matters. This involves identifying the underlying need, affected stakeholders and desired outcomes. The focus is not on technology, but on clarifying the challenge in its operational, social or strategic context. In this phase, governance processes are initiated by identifying governance requirements including privacy and ethical considerations. Automation Discovery Once the problem is defined, the discovery process assesses whether the challenge lends itself to automation. Progressing without this stage into Requirements Gathering risks designing a solution that may be unnecessary or unviable. Organisations consider a range of solution types including human-led, process-based, data-driven or AI-assisted approaches. 01 03 04 02 Framing and Feasibility Framing and Feasibility Design Build and Deploy Operate and Evaluate Major: New Requirements Minor fixes / improvements Stakeholder management Risk management Governance Design Build and Deploy Operate and Evaluate Problem definition Automation discovery Requirements gathering Organisation Readiness Delivery pathway Solution design and data assessment Implementation planning Data Processing Solution development Deployment & implementation Monitoring & maintenance Processing Impact evaluation System retirement Significant changes needed System Retirement
Consultation Draft for Review (December 2025). 87 Teams evaluate the high-level problem context to determine if automation could support the tasks. If automation is appropriate, they assess the high-level feasibility of candidate solutions, focusing on the nature of the challenge and the context in which it occurs. Stakeholders are engaged early to ensure that any operational/ethical concerns, including potential unintended consequences, are addressed. The outcome of this stage is a clear, consultable rationale showing how the decision on the type of automation aligns with organisational values, stakeholder needs, and responsible innovation. Requirements Gathering Organisations define what a proposed solution must achieve. Stakeholders are engaged to surface the full range of requirements before any design, procurement or build decisions are made. These are statements of intent, shaped by problem context and people affected. Examples may include: • Functional requirements describing what the solution must do. This could include improving data insights, predicting service demand or supporting AI-assisted drafting and streamlining of internal workflows. • Operational requirements define how the solution fits into existing systems and processes. For example, a small charity may need the solution to integrate with its donor database while a local authority may require solution integration with case handling software. • Ethical and governance requirements consider transparency, auditability, and human oversight needs. For example, if automated outputs need to be edited or reviewed before action is taken. This could relate to an AI assisted email system or a solution that leads to a funding decision. • Legal and compliance requirements consider what constraints must be adhered to. A multinational business will need to comply with all data protection laws across all its operational jurisdictions. • Technical and organisational requirements describe the conditions that must be met for the solution to work in practice. This could include language support across multiple regions, infrastructure compatibility, resource capacity and training needs. • Sustainability and scalability requirements describe the need for a solution to remain relevant as needs evolve, to grow alongside the organisation or sector, and to account for dependencies that might limit future adaptability. The outcome of this stage is a clear list of requirements in accessible, consultable language to ensure solutions are grounded in real world needs and are aligned with organisational values and responsibilities. Organisational Readiness Assessment Before selecting a delivery pathway, organisations consider their own readiness to deliver the requirements. This assessment ensures that the technical, operational, cultural and governance conditions are fit for purpose as a measure of preparedness. The assessment explores whether the organisation has the skills to adopt, build, or manage the solution, whether existing systems and data environments can support it, and whether teams and stakeholders are ready to adapt behaviours, roles, and ways of working. The process helps surface gaps that might otherwise undermine delivery and provides a structured basis for deciding whether to proceed. If the outcome is a "no-go", it signals what must change for the organisation to become ready. Delivery Pathway Selection
Consultation Draft for Review (December 2025). 88 Organisations assess how best to meet requirements whether by building in-house, commissioning a bespoke solution or procuring an existing product. This is a strategic decision that considers requirements, feasibility, cost, governance and organisational capacity. In this stage, organisations also consider data requirements in principle. They ask if the solution type is likely to depend on data they hold, whether the data is sensitive or regulated, or whether build or buy would affect how the data is accessed, processed and governed. The outcome of this stage is a chosen pathway that supports responsible implementation and reflects that organisation’s values, and responsibilities in responsible innovation. Solution Design and Data Assessment Organisations clarify how the solution will meet its functional goals and how it will integrate with existing systems and how it will uphold ethical and operational safeguards. For a build/commission model, a structured data assessment is carried out to identify the data needed, where it will come from, how it will be accessed and used. The data is evaluated for completeness, quality and relevance. Risks related to privacy, consent, bias and compliance are also assessed and it is confirmed if the data can be used legally and ethically within the chosen solution model. In builds or bespoke commissions, technical specifications are developed in this stage that consider future proofing and sustainability. This can include system architecture diagrams with data flows and integration points, as well as the establishment of performance criteria. The outcome of this stage is a design decision that is viable, ethical and legal, grounded in real-world constraints and responsible data use before technical specifications and implementation begin. Implementation Planning This stage prepares the organisation for ensuring the solution can be delivered responsibly, and whether this can be best achieved through internal development, external commissioning or configuration of an existing product. Implementation planning is when organisations define how the solution will be integrated into the existing environments, what dependencies must be managed and what governance measures need to be in place. This includes identifying affected stakeholders and clarifying what will change and why. As part of change management, a plan to support transition will be developed complete with training plan, role adaptation, and embedded feedback mechanisms. Organisations define long term viability including how the solution will be maintained, updated and governed over time. Ongoing support and resourcing are considered here as well as exit strategies and decommissioning pathways. Data Processing For a build/commission scenario this phase entails transforming the data for the next phase in development. This typically requires cleaning, handling missing value, normalising and selecting features that will support the model solution and testing. Solution Development In commission or build scenarios, this stage involves creating and configuring the technology in line with defined requirements. Development activities may include engineering , performance optimisation, interface design, and validation. Early iterations often take the form of proof-of-concept exercises, used to validate feasibility and demonstrate alignment with stakeholders needs. These may progress into pilot
Consultation Draft for Review (December 2025). 89 deployments, which test the solution in controlled, real-world conditions before full implementation. Teams work iteratively to ensure the solution remains aligned with its intended purpose and responsive to stakeholder feedback. Throughout this stage, documentation, testing, and traceability are maintained to support transparency, auditability, and future adaptation. Deployment and Implementation The solution has been deployed to an operational environment. This stage involves activating the solution and integrating into existing workflows. Implementation includes final setup and onboarding, providing the appropriate training, change management to support adoption, and ensure safeguards are effective, governance protocols are being followed, and support structures are adequate. During this phase, teams monitor solution behaviour to monitor performance and resolve issues. Monitoring and Maintenance This phase takes place post-deployment, and at this stage the performance of data-driven solutions and AI systems are continually monitored to identify deviations from its indented performance and ensure that the system meets its operational requirements. This may include updating components and retraining models with new data, managing data quality and compliance, and ensuring system accuracy, security and reliability. Impact Evaluation This phase involves assessing the real-world impact of the solution and evlautaing whether the intended outcomes have been met. It supports decisions to continue, adapt or retire the solution. System Retirement Systems may be retired when they become obsolete or need ot be replaced by more effective or contemporary alternatives. The retirement phase entails assessing and documenting the reasons for taking the AI system out of operation, notifying stakeholders and supporting them through the transition, shutting down the infrastructure, revoking access to data and archiving or disposing of data in compliance with best practices and regulatory requirements. Governance Governance across the AI lifecycle ensure that data is managed responsibly, ethically and in line with regulatory requirements. It involves establishing structures, practices and safeguards that ensure the development and use of AI is safe, ethical and responsible. This may include setting policies and procedures for data storage, data quality assessment, data processing, data security and data disposal. The plan should also include processes to ensure transparency and explainability in the development and deployment of AI systems. Governance is a building block to enable a responsible and ethical adoption of AI that remains fit for purpose throughout its deployment. The objective of governance is to ensure responsibility, accountability and transparency by having structures, practices and safeguards in place. It adapts to the evolution of risks, changes in context and solution maturity. Governance is not limited to oversight. It includes the following: • Capturing key decisions, assumptions and changes to support audit, review and confidence. • Authority is distributed appropriately so that concerns can be raised, escalated and