The AIRI Framework
The AIRI Framework provides a thorough, evidence-based system for measuring and developing AI readiness, at both individual and organisational levels. Built on 8+ years of real-world AI transformation experience, it uses behavioural anchors with measurable evidence, not self-assessed attitudes.
The target distribution. Most individuals and organisations should target L2 (AI Ready). Maturity level should be driven by business objectives. An organisation at L2 with strong governance and clear ROI is better positioned than one at L3 with governance gaps.
Two assessments, one architecture
Individually, each assessment provides practical insights. Together, they reveal critical alignment gaps between workforce readiness and organisational capability. We call this The AIRI Framework or just AIRI for short.
For Organisations
Organisational AI Readiness Index
Measures your organisation’s AI readiness across leadership, culture, governance, data readiness, and infrastructure. Used by governments, industry associations, and enterprises to benchmark capability and target interventions.
15 minutes · 15 questions · Free
Learn about oAIRI →For Individuals
Personal AI Readiness Index
Measures your individual AI readiness across mindset, ethics awareness, value creation, data literacy, and tools. The same proven framework, now applied to the individual. pAIRI scores feed directly into your organisation’s oAIRI scores.
15 minutes · 15 questions · Free
Learn about pAIRI →Shared architecture
Each dimension maps to one question in each assessment. The oAIRI name shows the organisational perspective, the pAIRI name the individual. Same architecture, different lens. 15 questions, about 15 minutes per assessment.
P1
pAIRI · Mindset
Is there commitment and capability to succeed?
D1 Management Support
Understanding AI
D2 AI Literacy
Learning Agility
D3 AI Talent
Role Positioning
D4 Workforce Adaptation
Delegation Disposition
D5 Experimentation Culture
Initiative & Experimentation
P2
pAIRI · Ethics & Responsibility
Are risks managed and ethics considered?
D6 AI Governance
Ethics Awareness
D7 AI Risk Control
Risk Awareness
D8 GenAI & Agent Risk Management
GenAI Critical Thinking
P3
pAIRI · Value Creation
Can value be identified and captured?
D9 Business Use Case
Use Case Identification
D10 GenAI Value Realisation
GenAI Productivity
P4
pAIRI · Data Literacy
Is data accessible and of sufficient quality?
D11 Data Quality
Data Evaluation
D12 Reference Data
Data Proficiency
P5
pAIRI · Tools & Technical Skills
Are tools, infrastructure, and standards in place?
D13 AI/ML/Data Infrastructure
Tool & Agent Proficiency
D14 Specification Standards & Practices
Specification Thinking
D15 GenAI & Agent Deployment
AI-Assisted Creation & Orchestration
Five pillars, fifteen dimensions, one shared architecture, the organisational and personal lenses on the same readiness model.
Five maturity levels
Both assessments map to the same five maturity levels. The AIRI Framework defines Level 2 (AI Ready) as the success target, for individuals and organisations alike. You do not need to become an AI expert. You need to use AI effectively.
L0
Range: 0.00, 0.99 · Anchor point: 0.00
Not started yet, and that is okay. Everyone starts here.
No meaningful AI engagement. Whether the subject is an individual who has never tried AI tools or an organisation with no AI strategy, everyone starts here. The key message: awareness is the first step, and starting now is what matters.
L1
Range: 1.00, 1.99 · Anchor point: 1.25
Knows AI exists but not yet using it consistently.
Aware of AI’s potential and actively exploring. For an individual, this means experimenting with AI tools and building foundational understanding. For an organisation, it means leadership sees the opportunity and initial pilots are underway. Curiosity is high, but systematic adoption has not started.
L2
Range: 2.00, 2.99 · Anchor point: 2.50
Success target
Uses AI effectively in work with measurable impact.
AI is actively used with measurable results. For individuals, this means effective daily use of AI tools with clear impact on productivity. For organisations, it means deployed solutions with governance in place. This is where the strategic question shifts from "what can we automate?" to "how should we restructure work?" AI Ready is the goal for most people and most organisations. You do not need to reach L4.
L3
Range: 3.00, 3.99 · Anchor point: 3.75
Beyond automation: actively restructuring how work gets done.
Beyond automation: actively building AI capabilities and restructuring how work gets done. Individuals at this level lead AI initiatives, mentor others, and build solutions. Organisations have dedicated AI teams, formal ethics processes, and transformation roadmaps. The focus shifts from adopting AI to reshaping work around it.
L4
Range: 4.00, 5.00 · Anchor point: 5.00
AI agents handle production; humans provide judgement.
Driving AI transformation at scale. AI agents handle routine production; humans provide judgement, strategy, and oversight. Individuals at this level architect AI ecosystems and influence industry standards. Organisations have mature governance at board level and continuous innovation embedded in operations.
The anchor point is the fixed score every answer at that level contributes, averaging these produces your pillar and overall scores.
Scoring
Each question uses behavioural anchors, descriptions of observable behaviour, not opinions. You pick the one that honestly matches where you are today. Each anchor maps to a fixed score, and pillar scores are simple averages of their questions.
P1 = (D1 + D2 + D3 + D4 + D5) / 5
P2 = (D6 + D7 + D8) / 3
P3 = (D9 + D10) / 2
P4 = (D11 + D12) / 2
P5 = (D13 + D14 + D15) / 3
Overall = (D1 + D2 + … + D15) / 15
Overall Level = min( band(Overall), band(P2) )The overall score weights all questions equally, not pillar averages. P1 (5 questions) naturally carries more weight than P3 or P4 (2 questions each).
Your overall level can never exceed your P2 level, your maturity cannot outgrow your ethics. P2 at L0 caps you at L0 (STOP), P2 at L1 caps you at L1 (CONDITIONS), P2 at L2+ lifts the cap (GO), the same three gate states as the Decision Engine.
pAIRI · individual
15 answers totalling 28.75 → 28.75 / 15 = 1.92 → band L1. P2 pillar = 1.67 → L1. min(L1, L1) → L1 (AI Aware).
oAIRI · leadership workshop
Question-weighted mean ≈ 1.29 → band L1. P2 pillar = 0.80 → L0 → STOP. min(L1, L0) → L0 (AI Unaware). Build the ethics foundation first.
oAIRI · strong org, thin ethics
Mean = 2.60 → band L2. P2 pillar = 1.50 → L1 → CONDITIONS. min(L2, L1) → L1 (AI Aware), capped from L2.
If your workforce’s average pAIRI is L2 but organisational Data Foundation (P4) is at L0, individual skills are running ahead of organisational infrastructure. This gap analysis is the foundation of effective AI transformation planning.
From assessment to action
Assess
15 questions, ~15 minutes. Teaching content between pillar groups. You learn while you assess.
Act
Your lowest pillars are your roadmap. The behavioural anchors at the next level describe, in observable behaviour, exactly what that level looks like. That is what you work toward.
Transform
Reassess every 3–6 months to track growth. oAIRI guides project prioritisation through the Decision Engine.
Level 2 marks the key shift from “what can we automate?” to “how should we restructure work around AI?” If you use AI to speed up existing workflows, that is automation. If you have eliminated a workflow entirely because AI handles it end to end, with human oversight at checkpoints, that is transformation. Both are valuable. The framework helps you know which stage you are in.
Why this framework
Every response option describes observable, measurable behaviour, not opinions or self-ratings. This prevents grade inflation and the Dunning-Kruger effect.
Developed from AI Singapore’s national programmes reaching 350,000+ professionals, 1,000+ organisations and deploying more than 350 real-world AI projects. Not theoretical, built from patterns observed in real AI adoption.
The assessment itself is a learning experience. Educational content between question groups means every participant leaves knowing more than when they started.
Version 3.2 adds the ethics level gate, your overall level is capped at your P2 level. Version 3.1 brought AI agent delegation proficiency, specification thinking, and trust calibration, reflecting how AI is actually used today.
Get the complete framework
Download the full framework document, 15 dimensions, scoring methodology, and behavioural anchors for both oAIRI and pAIRI, released under Creative Commons CC BY 4.0. Sign up to receive the download links by email.