All 20 Pillars
Original framework © Paul Gibbons · Adaptive Adoption™ · Layer 1 self-report · v0.1
AI Workforce Readiness Assessment
20 pillars · what your people do, not what they think they can do
Failure modes of the AI readiness assessment
Most AI readiness assessments are digital readiness surveys with the word “AI” substituted in. They inventory infrastructure, data quality, security posture and whether a change programme exists — then ask people to rate their own confidence with the tools. Every one of those things can be in place while nobody in the building has changed how they work. Three findings show how wide that gap runs:
- 78 per cent of organisations score themselves high on responsible-AI readiness. 14 per cent report having put it into operational practice.Both numbers are C-suite self-report — the survey audits what executives say about plans, not whether behaviour moved.(Accenture & Stanford, 2024)This instrument was built the other way round: not to collect claims about plans, but to test whether behaviour has changed.
- As much as a 39-point gap between what people think they accomplish with AI and what they actually do. In a randomised trial, experienced developers using AI tools were 19 per cent slower — and reported being 20 per cent faster. Every readiness survey asks the question that trial showed people cannot answer.(METR randomised controlled trial, 2025)
- A persistent, documented chasm between surveyed thoughts and actual behaviour. Fifty years of research puts the link between stated values or attitudes and what people subsequently do at a correlation of roughly .3 to .4 (Kraus, 1995 meta-analysis) — attitudes explain a fraction of conduct. A survey of how people feel about AI is a survey of how they feel. It is not a measure of what they do.
What this one does
This one measures behaviour. Not “are your people confident with AI” but how often, how much, how recently — enacted conduct you can observe. It is an AI-specific methodology, not a renamed technology survey, and it looks where the returns are actually decided: culture, change, governance and leaders. Twenty pillars across three domains, each read at up to three epistemic levels — what people report, what artefacts exist, and what telemetry and observation show. The distance between the first and the third is the finding no other assessment can produce, because no other assessment collects the third.
How you can use it
Test-drive it
The short version is free, takes a few minutes, and gives you a result on screen immediately — and a fuller readout by email.
Team workshop
One day. Your team takes the full instrument together and leaves with its own profile and three things to change.
C-suite workshop
The leadership layer read directly — where the top team stands, and where its own conduct is the constraint.
Enterprise assessment
Qualitative and quantitative, behavioural throughout, across a defined population: instrument, interviews, artefact audit and telemetry, with a board-ready readout. Scoped to your organisation.
AI WORKFORCE READINESS ASSESSMENT
Twenty pillars across three domains. For each pillar, select the description that best matches your organisation today — not aspirationally, but as enacted behaviour you can actually observe.
Layer 1 self-report. The full assessment adds evidence audit (Layer 2) and behavioural observation (Layer 3). A gap of 1.5+ between Layer 1 and Layer 3 is a reliable indicator of policy theatre.
© Paul Gibbons · AI Workforce Readiness Assessment · Adaptive Adoption™ · v0.1 · Self-report layer only · Full three-layer assessment available via facilitated engagement
Are your people onboard?
Are your workflows working?
Are your pilots scaling?
Three domains, twenty pillars
Change Agility™ · 7 pillars
- Master the Craft
- Embrace Complexity
- Consciously Manage Trust
- Put People First
- Design and Prototype
- Prioritise Behavior
- Manage Ethics Always
Leadership Delta™ · 7 pillars
- Strategic Imagination
- Active Modeling
- Creative Cultivation
- Friction Courage
- Trust Calibration
- Ethical Stewardship
- Systems Orchestration
Behavioral Governance™ · 6 pillars
- Decision Rights
- Agent Authority
- Risk Intelligence
- Governance Intelligence
- 1st-Derivative Talent
- Strategic Coherence
What does an AI readiness assessment measure?
Most measure technical preconditions — infrastructure, data quality, security posture, whether a change programme exists — plus people’s self-rated confidence with the tools. This one measures behaviour: how often, how much, how recently people actually use AI, read across twenty pillars in three domains — organisational practice (Change Agility™), leadership behaviour (Leadership Delta™), and enabling guardrails (Behavioral Governance™) — because culture, change, governance and leaders are where the returns on AI are actually decided.
Is this a self-assessment?
The free short version is a structured self-report (Layer 1), scored instantly on screen. The full assessment adds an evidence audit (Layer 2) and behavioural observation and telemetry (Layer 3). The distance between what people report and what observation shows is the finding no survey can produce — a gap of 1.5+ between Layer 1 and Layer 3 is a reliable indicator of policy theatre.
How is this different from a digital readiness survey?
Most AI readiness assessments are digital readiness surveys with the word “AI” substituted in: they audit whether a plan exists, not whether behaviour moved. Fifty years of research puts the link between stated attitudes and subsequent conduct at a correlation of roughly .3 to .4 (Kraus, 1995 meta-analysis) — a survey of how people feel about AI is a survey of how they feel. This is an AI-specific methodology that reads enacted conduct you can observe, which is why it can see the gap between the 78% of organisations that score high on responsible-AI readiness and the 14% that have operationalised it (Accenture & Stanford, 2024).
People-First AI™, a term coined by Paul Gibbons with James Healy in 2024, holds that the binding constraint on artificial intelligence in organisations is human and organisational rather than technological — and that designing for people is not the ethical alternative to performance but the mechanism that produces it.