SFIA skills for AI and machine learning

The SFIA skill Strategic planning defines the professional capability to develop and maintain organisational strategies that align plans, actions and resources with business objectives. AI can be incorporated alongside other strategic capabilities and investments rather than treated as an objective in itself.

The SFIA skill Enterprise and business architecture defines the professional capability to translate organisational strategy into coordinated operating models, capability changes and roadmaps. It contributes to AI change spanning business, data, applications and infrastructure.

The SFIA skill Business situation analysis defines the professional capability to investigate business situations and recommend improvement action. It helps separate the underlying need from assumptions about AI as the solution.

The SFIA skill Feasibility assessment defines the professional capability to define and evaluate change options. It allows AI approaches to be compared with alternative interventions and assessed against organisational constraints.

The SFIA skill Emerging technology monitoring defines the professional capability to identify and assess emerging technologies, products, services, methods and techniques. It supports systematic assessment of AI developments and their possible organisational implications.

The SFIA skill Innovation management defines the professional capability to identify, prioritise, incubate and exploit opportunities from digital technologies. It supports experimentation while connecting AI opportunities to organisational objectives and decision-making.

The SFIA skill Investment appraisal defines the professional capability to assess possible investments and projects. It can support decisions about AI platforms, products, infrastructure, development and transformation initiatives.

The SFIA skill Portfolio management defines the professional capability to select, prioritise and balance programmes, projects and ongoing services as a coherent portfolio. It enables AI initiatives to be considered alongside other organisational investments rather than through a separate AI-specific investment process.

The SFIA skill Benefits management defines the professional capability to forecast, plan and monitor expected benefits. It helps organisations test whether AI initiatives actually produce the outcomes used to justify them.

The SFIA skill Product management defines the professional capability to manage products and services across their lifecycle. It supports decisions about incorporating, evolving or withdrawing AI capabilities based on evidence and changing needs.

The SFIA skill Budgeting and forecasting defines the professional capability to develop budgets and forecasts using scenario planning, sensitivity analysis and risk assessment. This supports planning for variable and consumption-based AI expenditure.

The SFIA skill Cost management defines the professional capability to monitor, analyse and control costs and identify optimisation opportunities. It supports decisions about consumption, model choice and other cost-performance trade-offs.

The SFIA skill Financial analysis defines the professional capability to analyse financial information, develop models and conduct cost-benefit, scenario and sensitivity analysis to support decision-making.

The SFIA skill Sustainability directly addresses policies, standards, metrics and initiatives that reduce the environmental impact of technology.

The SFIA skill Machine learning directly requires practitioners to anticipate sustainability implications when developing, deploying and managing machine-learning systems.

The SFIA skill Measurement contributes defined measures, targets, thresholds and evidence for energy, carbon, water, hardware use and other environmental impacts.

The SFIA skill Solution architecture contributes evaluation of environmental requirements and trade-offs across models, data, applications, infrastructure and external services.

The SFIA skill Infrastructure design contributes choices about compute, storage, networking, resilience and resource efficiency that shape AI's environmental impact.

The SFIA skill Capacity management contributes forecasting and optimisation of processing and infrastructure capacity against workload, service, cost and resource constraints.

The SFIA skill Asset management contributes lifecycle decisions about acquisition, utilisation, maintenance, reuse and disposal of AI-related hardware and services.

The SFIA skill Sourcing contributes environmental criteria to make-or-buy decisions, supplier selection, procurement requirements and evaluation of externally supplied AI services.

The SFIA skill Governance defines the professional capability to establish and operate frameworks for decision-making, legitimate authority and accountability. It can define who may authorise AI uses, what may be delegated and where human responsibility remains.

The SFIA skill Risk management defines the professional capability to establish systematic approaches to organisational risk. AI risks can therefore be managed alongside other business, operational and technology risks.

The SFIA skill Artificial intelligence (AI) and data ethics directly addresses the professional capability to apply ethical principles throughout the design, development, deployment and use of AI and data technologies.

The SFIA skill Information and data compliance defines the professional capability to implement and promote compliance with information and data management legislation.

The SFIA skill Audit defines the professional capability to conduct independent, risk-based assessments of processes, controls and compliance. It can provide assurance over AI governance arrangements and supporting evidence.

The SFIA skill Stakeholder relationship management defines the professional capability to analyse, manage and influence stakeholder relationships systematically. It contributes where AI decisions have benefits, risks and consequences across organisational boundaries.

The SFIA skill Artificial intelligence (AI) and data ethics directly addresses ethical impact assessment and the wider consequences of AI and data technologies.

The SFIA skill Governance defines the professional capability to establish authority, decision rights and oversight for commissioning, approving and reviewing impact assessments.

The SFIA skill Risk management defines the professional capability to identify, assess, treat and monitor risks revealed through impact assessment.

The SFIA skill Stakeholder relationship management contributes structured engagement with users, affected groups and other stakeholders when identifying impacts and agreeing responses.

The SFIA skill Audit contributes independent, risk-based assessment of impact-assessment processes, supporting evidence, controls and compliance.

The SFIA skill Governance defines the professional capability to establish and operate decision-making, authority and accountability arrangements. It provides the primary capability for bringing actual AI use within approved oversight and control.

The SFIA skill Information security defines the professional capability to establish and operate security controls. It contributes to identifying and controlling data exposure, unauthorised services and other security risks from unapproved AI use.

The SFIA skill Asset management defines the professional capability to manage technology assets and explicitly addresses unauthorised assets, including cloud services. It contributes visibility and control of unmanaged AI tools and services.

The SFIA skill Risk management defines the professional capability to identify, assess and treat organisational risk. It supports proportionate decisions about unapproved or unmanaged AI use.

The SFIA skill Sourcing defines the professional capability to procure or commission products and services. It provides controlled routes for meeting legitimate AI needs identified through shadow-AI discovery.

The SFIA skill Business process improvement defines the professional capability to analyse and redesign business processes and exploit new technologies and automation. It supports changes to workflows and operating practices where AI alters how business outcomes are achieved.

The SFIA skill Requirements definition and management defines the professional capability to establish and manage requirements throughout delivery and operation. AI-specific requirements can be managed alongside functional, service, security, regulatory and other requirements.

The SFIA skill User research defines the professional capability to investigate user behaviour, needs and motivations. This helps ground AI design in real contexts of use rather than assumed user behaviour.

The SFIA skill User experience analysis defines the professional capability to understand context of use and specify user experience requirements and goals. This can include how AI recommendations, generated content and uncertainty are presented and used.

The SFIA skill User experience design defines the professional capability to design interactions and experiences for systems, products and services. It can address how people interact with AI capabilities within wider work and decision processes.

The SFIA skill User experience evaluation defines the professional capability to validate products and services against user experience goals and measures. It can identify usability problems and inappropriate patterns of reliance on AI.

The SFIA skill Artificial intelligence (AI) and data ethics directly addresses transparency, accountability and the ethical consequences of AI-enabled decisions and actions.

The SFIA skill User experience design defines the professional capability to design usable and accessible explanations, disclosures, intervention controls and routes for challenge.

The SFIA skill User experience analysis contributes understanding of each audience's context, tasks and information needs when defining explanation and contestability requirements.

The SFIA skill User research contributes evidence about how users and affected people understand AI information, experience decisions and seek review or correction.

The SFIA skill User experience evaluation contributes validation that explanations and challenge routes are understandable, accessible and usable in realistic contexts.

The SFIA skill Governance defines the professional capability to establish decision rights, human review authority, escalation routes and accountability for organisational responses.

The SFIA skill Stakeholder relationship management contributes planned engagement and communication with users, affected groups, procurers, regulators and other interested parties.

The SFIA skill Data management defines the professional capability to establish policies and practices that control, protect and optimise data assets. It provides the broader stewardship needed around AI data.

The SFIA skill Data engineering defines the professional capability to design, build, operationalise, secure and monitor data pipelines and stores. It supports dependable movement and preparation of data for AI workloads.

The SFIA skill Data modelling and design defines the professional capability to represent, communicate and manage data requirements and assets. It supports consistent interpretation and integration of AI-related data.

The SFIA skill Information management defines the professional capability to enable effective management and use of information assets. It supports responsibilities, access, lifecycle, protection and use beyond individual AI applications.

The SFIA skill Analytical classification and coding defines the professional capability to interpret information and assign classifications or labels using domain knowledge, standards and guidance. This includes data labelling for machine learning where appropriate.

The SFIA skill Data analytics defines the professional capability to extract, analyse and communicate insights from structured and unstructured data. It can support AI development, evaluation and operational decision-making.

The SFIA skill Machine learning directly addresses the professional capability to develop systems that learn from data and experience.

The SFIA skill Data science defines the professional capability to apply mathematics, statistics, data mining and predictive modelling to generate insight and value from data.

The SFIA skill Numerical analysis defines the professional capability to create, analyse, implement, test and improve algorithms for numerically solving mathematical problems. It applies where the underlying AI work requires this specialist capability.

The SFIA skill High-performance computing defines the professional capability to use advanced computing systems and specialised programming techniques for complex computational problems. It is relevant where AI workloads require this scale of computation.

The SFIA skill Methods and tools defines the professional capability to lead the adoption, management and optimisation of methods and tools in alignment with organisational objectives.

The SFIA skill Formal research defines the professional capability to create new knowledge systematically through experimentation, evaluation and dissemination. It distinguishes formal research from routine investigation or development.

The SFIA skill Solution architecture defines the professional capability to develop and communicate multidimensional solution architectures and make appropriate architectural trade-offs.

The SFIA skill Systems design defines the professional capability to design systems against requirements and agreed architectures. It supports the integration of AI components with wider system behaviour.

The SFIA skill Software design defines the professional capability to architect and design software against specified requirements, standards and principles.

The SFIA skill Programming/software development defines the professional capability to develop software components that deliver stakeholder value. AI-specific libraries and techniques sit within this wider development capability.

The SFIA skill Systems integration and build defines the professional capability to plan, implement and control the integration of system elements, subsystems and interfaces.

The SFIA skill Configuration management defines the professional capability to identify, control, account for and audit configuration items and their relationships. It can provide traceability across the components of AI-enabled systems.

The SFIA skill Infrastructure design defines the professional capability to design technology infrastructure to meet business and technical requirements.

The SFIA skill Systems and software lifecycle engineering defines the professional capability to establish repeatable lifecycle practices across systems and software. AI-enabled systems may require these practices to cover models, data and other evolving components.

The SFIA skill Configuration management directly addresses the identification, status, version, relationships and lifecycle traceability of configuration items.

The SFIA skill Machine learning directly addresses traceability for outcomes produced by machine-learning systems and evidence about model development, evaluation and operation.

The SFIA skill Data management contributes governance of data assets, metadata, quality, protection, retention and provenance used throughout the AI lifecycle.

The SFIA skill Requirements definition and management contributes controlled evidence of intended purpose, requirements, constraints, acceptance criteria and their lifecycle changes.

The SFIA skill Functional testing contributes evidence about specified and expected system behaviour, test conditions, results, defects and corrective action.

The SFIA skill Non-functional testing contributes evidence about performance, security, scalability, robustness and other required system qualities.

The SFIA skill Quality assurance contributes evidence that defined quality requirements, practices and controls are followed and organisational quality objectives are met.

The SFIA skill Records management contributes controlled retention, retrieval and disposal of evidence that must remain available beyond active configuration management.

The SFIA skill Audit contributes independent examination of evidence, traceability, controls and compliance against defined criteria.

The SFIA skill Safety engineering defines the professional capability to apply safety principles throughout the lifecycle of AI used in safety-related systems, including hazard analysis, requirements, verification and safety-case preparation.

The SFIA skill Hardware design defines the professional capability to design AI hardware that meets processing, power, environmental, reliability and device-integration constraints.

The SFIA skill Real-time/embedded systems development defines the professional capability to develop AI software that operates under timing, power, reliability, safety and physical-device constraints.

The SFIA skill Safety assessment defines the professional capability to independently assess AI used in safety-related systems, including standards, safety-integrity requirements and risk-reduction evidence.

The SFIA skill Functional testing defines the professional capability to investigate systems and services against specified or unspecified functional requirements and characteristics.

The SFIA skill Non-functional testing defines the professional capability to evaluate non-functional characteristics of systems and services. AI-specific measures and evaluation methods can be applied where relevant requirements have been established.

The SFIA skill User acceptance testing defines the professional capability to validate systems, products, business processes and services against defined acceptance criteria.

The SFIA skill Quality management defines the professional capability to establish and operate organisational quality frameworks, standards, measures and improvement practices. It complements testing and quality assurance by addressing the management system within which AI quality is governed and improved.

The SFIA skill Quality assurance defines the professional capability to assess whether organisational quality objectives are being met through ongoing and periodic review.

The SFIA skill Measurement defines the professional capability to develop and operate measurement arrangements that meet organisational information needs. It can support model, system, service and outcome measures.

The SFIA skill Service acceptance defines the professional capability to define and assess service acceptance criteria covering functionality, operational support, performance, safety, security and quality.

The SFIA skill Artificial intelligence (AI) and data ethics contributes assessment and treatment of harmful, unfair or irresponsible AI behaviour and its consequences.

The SFIA skill Requirements definition and management contributes explicit requirements for permitted behaviour, human oversight, action boundaries, failure tolerances and acceptance.

The SFIA skill Machine learning directly addresses model performance, robustness, fairness, bias, issue diagnosis and adaptation before and after deployment.

The SFIA skill Functional testing contributes evaluation of specified and expected behaviours, including safe responses, refusal, escalation and tool-execution logic.

The SFIA skill Non-functional testing contributes evaluation of robustness, security and other safety-related qualities under realistic and adverse conditions.

The SFIA skill Information security defines the professional capability to establish controls and embed secure-by-design practices against manipulation, disclosure and unsafe action.

The SFIA skill Penetration testing contributes adversarial evaluation of prompt injection, control bypass, data exfiltration and unsafe use of connected tools or services.

The SFIA skill Identity and access management contributes control of non-human identities, privileges, delegated authority, tool permissions and access boundaries.

The SFIA skill User experience design contributes effective human intervention, override, confirmation and escalation within interactions with generative and agentic systems.

The SFIA skill Information security defines the professional capability to define and operate frameworks of security controls and security management strategies. AI-specific threats can be incorporated into these arrangements.

The SFIA skill Information assurance defines the professional capability to protect against and manage risks associated with information and information systems.

The SFIA skill Security operations defines the professional capability to manage and administer operational security measures using tools and intelligence.

The SFIA skill Vulnerability assessment defines the professional capability to identify and classify security vulnerabilities and mitigate or eliminate their impact.

The SFIA skill Penetration testing defines the professional capability to test security controls by emulating the tools and techniques of likely attackers. AI-specific attack techniques can be incorporated where relevant.

The SFIA skill Identity and access management defines the professional capability to manage identity verification and access permissions within organisational systems and environments. The AI context may require extension of these practices to increasingly complex non-human actors and services.

The SFIA skill Release management defines the professional capability to manage releases into production while addressing business objectives and compliance requirements. Models and related components can be treated as controlled release artefacts.

The SFIA skill Deployment defines the professional capability to transition software into live usage while managing risks and ensuring it operates as intended.

The SFIA skill Infrastructure operations defines the professional capability to provision, configure, operate and optimise technology infrastructure across physical, virtual and cloud environments.

The SFIA skill Incident management defines the professional capability to coordinate responses to incidents, minimise negative impacts and restore services.

The SFIA skill Problem management defines the professional capability to manage problems throughout their lifecycle and address causes rather than repeatedly treating symptoms.

The SFIA skill Availability management defines the professional capability to ensure services achieve required levels of availability for current and future needs.

The SFIA skill Capacity management defines the professional capability to ensure service components have sufficient capacity and performance to meet current and planned needs.

The SFIA skill Service level management defines the professional capability to agree service targets and assess and manage service delivery against them.

The SFIA skill Application support defines the professional capability to support and maintain live applications. It can identify operational problems and improvement opportunities involving AI components.

The SFIA skill Change control defines the professional capability to assess risks associated with proposed changes and ensure that changes are controlled and coordinated.

The SFIA skill Continuity management defines the professional capability to assess business impacts, develop and test continuity arrangements and maintain required operations during disruption. It also covers continuity across externally supplied and cloud-based services and supply chains.

The SFIA skill Machine learning directly addresses MLOps, model monitoring, issue diagnosis, retraining, adaptation and evaluation of performance, robustness, fairness and bias.

The SFIA skill Measurement defines the professional capability to establish measures, targets, thresholds, data collection and analysis for operational AI behaviour and outcomes.

The SFIA skill Systems and software lifecycle engineering contributes the automated lifecycle environment, feedback practices and continual improvement needed to act on monitoring evidence.

The SFIA skill Application support contributes investigation and improvement of live AI-enabled applications as operational evidence, incidents and user needs change.

The SFIA skill Security operations contributes live monitoring and response for threats, misuse, anomalous behaviour and failures of operational security controls.

The SFIA skill Change control contributes risk assessment, authorisation and coordination when monitoring evidence leads to changes in models, prompts, data, software or services.

The SFIA skill Functional testing contributes repeatable checks of AI-enabled behaviour after changes or when monitoring identifies possible functional deterioration.

The SFIA skill Non-functional testing contributes continuous assessment of performance, robustness, security and other system qualities against defined expectations.

The SFIA skill Sourcing defines the professional capability to manage and advise on procurement and commissioning. AI-specific criteria can be considered alongside cost, service, risk and wider organisational requirements.

The SFIA skill Supplier management defines the professional capability to align supplier performance with sourcing strategies and organisational objectives.

The SFIA skill Contract management defines the professional capability to manage formal contractual obligations throughout the contract lifecycle.

The SFIA skill Asset management defines the professional capability to manage assets from acquisition and operation through maintenance and disposal.

The SFIA skill Specialist advice defines the professional capability to provide authoritative advice on technology-related intellectual property, rights, licensing and provenance.

The SFIA skill Asset management directly includes intellectual property, licences, agreements and other AI-related assets throughout acquisition, operation, maintenance and disposal.

The SFIA skill Contract management contributes negotiation and operation of enforceable rights, restrictions, warranties, indemnities, attribution and supplier obligations.

The SFIA skill Data management contributes governance of rights, restrictions, provenance, retention and permitted use for data and knowledge sources.

The SFIA skill Configuration management contributes versioned traceability between models, datasets, prompts, components, licences, approvals and deployed configurations.

The SFIA skill Sourcing contributes evaluation of supplier terms, licensing models, provenance evidence and rights-related risks during procurement and commissioning.

The SFIA skill Supplier management contributes continuing oversight of supplier compliance, changes, disclosures, disputes and risks affecting rights or licences.

The SFIA skill Project management defines the professional capability to deliver agreed project outcomes using management, collaboration, leadership and governance appropriate to the context.

The SFIA skill Programme management defines the professional capability to plan and coordinate related projects and activities in support of business objectives.

The SFIA skill Organisational change management defines the professional capability to plan, design and implement transition to a required future state.

The SFIA skill Organisation design and implementation defines the professional capability to design and implement integrated organisation structures and cultures. It can support redesign where AI materially alters operating models and responsibilities.

The SFIA skill Performance management defines the professional capability to improve individual and workgroup performance against agreed objectives and measurable results. Where AI changes work, objectives can reflect revised responsibilities and expected outcomes without rewarding AI use for its own sake.

The SFIA skill Workforce planning defines the professional capability to project future demand for people and skills and plan how those requirements will be met.

The SFIA skill Job analysis and design defines the professional capability to analyse and design jobs and role structures. It can distinguish AI-related responsibilities within existing roles from requirements for new specialist roles.

The SFIA skill Organisational capability development defines the professional capability to assess organisational capabilities, prioritise improvements and support their implementation.

The SFIA skill Competency assessment defines the professional capability to assess knowledge, skills, competency and behaviours against frameworks such as SFIA. It can assess evidence of professional judgement and responsible AI use as well as technical knowledge.

The SFIA skill Learning design and development defines the professional capability to design learning resources and interventions that transfer knowledge, develop skills and change behaviours.

The SFIA skill Competency assessment contributes evidence of current capability and gaps that learning interventions need to address.

The SFIA skill Job analysis and design contributes the role and responsibility context needed to define relevant learning outcomes for different workforce groups.

The SFIA skill Learning delivery defines the professional capability to transfer knowledge, develop skills and change behaviours using appropriate learning approaches.

The SFIA skill Professional development defines the professional capability to support development in line with individual career needs and organisational requirements.

The SFIA skill Knowledge management defines the professional capability to systematically capture, develop and use organisational knowledge. This can help organisations learn from AI experiments, implementations, incidents and emerging practices.

The SFIA skill Workforce planning defines the professional capability to project future demand for people and skills and plan the workforce supply needed to meet organisational requirements.

The SFIA skill Resourcing defines the professional capability to acquire, deploy and onboard resources. It contributes where AI capability must be recruited or obtained through contingent staffing.

The SFIA skill Sourcing defines the professional capability to manage or advise on procurement and commissioning. It contributes where AI capability is obtained through consultancy or externally supplied professional services.

The SFIA skill Supplier management contributes continuing management of external providers supplying AI-related capability and expertise.