Training activity
Employees given access to tools
Courses completed and certificates issued
Participant satisfaction and self-reported confidence
CFTE / AI Proficiency & Measurement
Define capability.Demonstrate impact.
Organisations can count AI tools deployed, employees trained and courses completed. The critical question is what people can actually do with AI, how their work is changing and which outcomes are improving. These are three connected forms of evidence, but they need different methods of measurement.
CFTE / Workforce Transformation at ScaleResearch and perspectives on AI proficiency, measurement and workforce capability
01 / The measurement problem
Thousands trained. But how many are capable?
Consider a bank that has deployed AI tools, trained thousands of colleagues and established internal champions. Those are useful milestones, but leadership still needs evidence of who can apply AI independently, verify outputs, recognise risks and make sound decisions in their role.
Demonstrated capability
Can employees select appropriate uses for AI?
Can they independently apply and verify outputs?
Can they use AI responsibly in real situations?
The central tension
Training completion provides evidence of participation. It does not, on its own, demonstrate proficiency or establish business impact.
02 / Define what we mean
Measure transferable proficiency, not just familiarity with today's tools.
CFTE's AI Proficiency Framework helps distinguish four related concepts. Tool fluency can change rapidly; proficiency is more transferable. Applied capability depends on the professional context, while readiness also depends on the role, risks and institutional environment.
The core foundation
AI proficiency
Transferable understanding, application and judgement that remain useful as tools evolve.
Current technology
Tool fluency
Familiarity with particular models, interfaces and applications, which changes over time.
Professional context
Applied capability
Ability to combine proficiency with domain knowledge and real-work decisions.
Overall judgement
AI readiness
Whether someone is prepared to use AI effectively in their role, risk environment and institutional setting.
03 / CFTE AI Proficiency Framework
Three public levels: increasing autonomy, judgement and responsibility.
The CFTE AI Proficiency Framework (April 2026) provides a shared language for describing AI capability across roles. It combines three public proficiency levels, ten domains and three assessment dimensions: knowledge, skills and behaviours.
LEVEL 1
AI Literacy
Safe, disciplined use in routine professional contexts; understanding limits and verifying outputs.
LEVEL 2
Applied AI Practitioner
Independent problem-solving, appropriate tool selection, output validation and workflow application.
LEVEL 3
AI Systems and Decision Leader
Systems thinking, orchestration, implementation judgement and governance responsibility.
KnowledgeWhat people understand
SkillsWhat people can do
BehavioursHow people exercise judgement and responsibility
AI foundationsApplications & use casesTools & methodsDataRisks & limitationsRegulation & ethicsImplementationStrategy & impactEmerging trendsTechnology landscape
The three public levels are the reporting view; the six developmental bands (0–5) provide more detailed calibration. AI Literacy aligns with Basic Proficiency (band 2), Applied AI Practitioner with Working Proficiency (band 3), and AI Systems and Decision Leader with Advanced Application and Strategic Mastery (bands 4–5).
04 / In professional practice
What credible evidence of AI proficiency looks like.
Consider a compliance professional assessing an AI-generated case summary. A knowledge question reveals whether they understand the risks, but applying AI responsibly also requires checking the evidence and deciding when to challenge or escalate a result.
Knowledge
Understand the limits
Recognise relevant AI risks, common failure modes and regulatory expectations.
Skills
Use and validate
Apply AI to a realistic analysis, check sources and correct weak outputs.
Behaviours
Exercise judgement
Protect sensitive information, recognise unsuitable uses and escalate uncertainty.
A sound assessment approach can combine knowledge checks, realistic scenarios, situational judgement and applied tasks. Each reveals a different part of capability; none should be mistaken for direct proof of improved business outcomes.
05 / The bigger picture
From measured proficiency to evidence of better work.
For business leaders, capability is the starting point. Useful evaluation also examines whether people apply their skills in real workflows and whether operational results improve. These are three linked questions, not one combined score.
01 / CAPABILITY
Can they do it?
Assess knowledge, skills and behaviours, and identify strengths and gaps for different workforce populations.
02 / APPLICATION
Has work changed?
Observe whether AI is used effectively in real workflows, decisions and responsibilities.
03 / BUSINESS IMPACT
What has improved?
Measure relevant changes in speed, quality, costs, customer experience, risk, capacity or revenue.
Interpreting the evidence: A diagnostic helps establish capability and development needs. Understanding business impact also requires evidence of changed work, relevant operational metrics and careful interpretation of other factors.
06 / Use the evidence
Turn measurement into better workforce decisions.
A common baseline makes workforce capability visible. Role-specific expectations then help leaders decide what to develop: a relationship manager, compliance officer, executive and technical specialist may need different levels and patterns of AI capability.
Baseline
Know where people are
Identify current capability levels and patterns across teams, roles and regions.
Development
Target what matters
Prioritise learning pathways, reinforcement and opportunities for applied practice.
Outcomes
Revisit the evidence
Track progress in proficiency and connect it to observed changes in work and measurable results.
Implications for decision-makers
What HR and business leaders can do now.
The same evidence supports different decisions about development, adoption and operational performance.
CHRO / Head of HR
Start with a role-based capability baseline for priority populations. Identify material gaps in knowledge, applied skills and responsible behaviours; target learning and practice where needed; then reassess to establish whether capability has improved.
COO / Business leader
Choose a high-value workflow and define expected benefits and control requirements. Observe how people apply AI, then compare quality, review time, exceptions and rework against a baseline. Separate improved capability from proven business impact.
Research and practical resources
Explore the framework, diagnostic and development pathways.
CFTE connects a public reference framework to assessment and role-based learning, helping institutions make more targeted decisions about AI workforce capability.
Three public levels, ten domains and knowledge, skills and behaviours.
Explore the framework ↗An assessment approach combining knowledge, skills and behavioural evidence to identify workforce capability gaps.
Explore the diagnostic ↗Translate common AI proficiency expectations into targeted learning and workforce development for different roles.
Explore workforce development ↗Continue the thinking
Explore the wider implications of AI.
Measuring AI proficiency connects to changing work, professional contribution and organisational performance. Explore the wider picture.
Connect with CFTE
What can your workforce demonstrate?
Identify where your organisation needs greater visibility into AI capability. CFTE can help define requirements, assess workforce strengths and gaps, and build development programmes connected to business priorities.