AI Readiness Assessment Checklist [2026]

An AI readiness assessment is the gap analysis you run before committing budget to AI tools, training, or transformation programs. Skip it, and you end up buying licenses nobody uses, writing policies nobody follows, and measuring progress with vibes instead of data. This checklist gives L&D leads, transformation directors, and engineering managers a structured, printable framework to evaluate where their organisation actually stands — across people, process, and infrastructure — before spending a dollar.

The checklist covers seven areas. Each one maps to a concrete organisational capability. Score them honestly, and you'll know exactly which gaps to close first and which vendor pitches to ignore.

What AI Readiness Actually Means

AI readiness is an organisation's capacity to adopt, integrate, and sustain AI tools and practices across its workforce and operations. It is not a technology question alone — it spans people skills, governance structures, cultural norms, and measurement systems.

Many teams conflate AI readiness with having purchased an enterprise ChatGPT licence. That's tooling, not readiness. A 2024 McKinsey Global Survey found that only 11% of organisations reported having the talent needed to fully capture value from generative AI, despite 72% reporting adoption of AI in at least one business function. The gap between "we have tools" and "we can use them effectively" is where readiness lives.

AISA defines AI readiness across five measurable dimensions: prompting and communication, critical thinking, technical understanding, workflow application, and safety and responsibility. These map directly to the skills individuals need to make AI tools productive rather than decorative. If your people can't prompt effectively, verify outputs, or integrate AI into existing workflows, your organisation isn't ready — regardless of how many seats you've provisioned.

Readiness is also distinct from AI fluency, which describes an individual's demonstrated skill level. Readiness is the organisational wrapper: do you have the policies, governance, culture, and measurement infrastructure to support fluency at scale? Think of fluency as the individual muscle; readiness is the gym, the programme, and the coach.

The 7-Point AI Readiness Assessment Checklist

This checklist is designed to be completed by a cross-functional team — ideally someone from L&D, IT/engineering, legal or compliance, and a business unit leader. Each item requires a concrete answer, not an aspiration.

1. People Skills Baseline

The single most important — and most frequently skipped — item on any AI readiness checklist. Before you train anyone, you need to know where they are now.

What to assess:

• Do you have a current, evidence-based measurement of AI fluency across your workforce?
• Can you segment skill levels by role, department, and seniority?
• Do you know the gap between self-reported confidence and actual demonstrated ability?

Why it matters: Self-assessment is unreliable. AISA data from 2,482 completed assessments shows an average prediction gap of 17.8 points — people predicted a composite score of 62.4 but actually scored 44.6. Without objective measurement, you'll design training for the wrong level.

Evidence required: A named assessment tool, completion rates by department, and a summary of score distributions.

2. AI Policy

Policy is the guardrails that make adoption safe. Without it, individuals make their own rules — and those rules will be inconsistent.

What to assess:

• Do you have a published, accessible AI acceptable-use policy?
• Does it cover data classification (what can and cannot be entered into AI tools), approved tools, and escalation paths?
• Has it been reviewed by legal in the last six months?
• Do employees know it exists?

Evidence required: A link to the policy document, the date of last review, and a metric on employee awareness (e.g., acknowledgement rate, training completion).

3. Tooling and Infrastructure

Tooling is the easiest box to tick and the least meaningful on its own. But you still need it.

What to assess:

• Which AI tools are officially sanctioned? List them.
• Are enterprise-grade versions deployed (with data privacy controls, SSO, audit logging)?
• Is there shadow AI usage — teams using unsanctioned tools with company data?
• Do you have API access for teams that need programmatic integration?

Evidence required: A tool inventory, licence counts vs. active usage rates, and a shadow-AI audit (even informal).

4. Governance and Oversight

Governance answers the question: who decides what, and who is accountable when something goes wrong?

What to assess:

• Is there a named AI governance owner (person or committee)?
• Do you have a decision framework for evaluating new AI use cases — including risk classification?
• Are human-in-the-loop requirements defined for high-stakes outputs?
• Is there a process for incident response when AI produces harmful, biased, or incorrect outputs?

The EU AI Act is now in phased enforcement. If you operate in or sell into the EU, governance isn't optional — it's a compliance requirement. UNESCO's 2024 Readiness Assessment Methodology identifies governance as one of five pillars, alongside infrastructure, legal frameworks, education, and economic capacity.

Evidence required: A governance charter or terms of reference, named owners, and at least one documented decision or incident review.

5. Measurement Framework

If you can't measure progress, you can't manage a transformation. This goes beyond "number of prompts sent."

What to assess:

• Do you have defined KPIs for AI adoption (not just usage, but impact)?
• Can you measure individual skill progression over time?
• Are you tracking automation ROI — time saved, error reduction, or throughput gains — at the team level?
• Is there a baseline measurement to compare against?

Evidence required: A measurement plan with named metrics, a baseline date, and a reporting cadence.

6. Culture and Change Readiness

Culture determines whether AI tools get used or gather dust. This is the hardest item to score because it's the most subjective — but there are concrete signals.

What to assess:

• Do leaders visibly use AI tools and talk about their own learning?
• Is there psychological safety to experiment and fail with AI?
• Are there AI champions — informal or formal — who help peers adopt tools?
• Is AI framed as augmentation ("this helps you do better work") or replacement ("this replaces your job")?

Evidence required: Leadership communications referencing AI use, an internal survey or pulse check on sentiment, and examples of peer-led adoption.

7. Budget and Resourcing

Readiness without budget is a wish list. This item forces specificity.

What to assess:

• Is there a dedicated budget line for AI tools, training, and assessment?
• Is the budget sized for the full adoption lifecycle (pilot → rollout → measurement → iteration), not just the pilot?
• Are there named people (not just committees) with time allocated to drive AI adoption?
• Have you accounted for ongoing costs — licence renewals, model upgrades, retraining?

Evidence required: A budget document or business case, named programme owners with allocated FTE, and a timeline.

How to Score Your AI Readiness Checklist

Each of the seven items should be scored on a simple three-level scale. The goal is honesty, not optimism. If you're unsure, score it lower — uncertainty is itself a signal.

Interpreting your total (0–14):

A few scoring principles:

• Score the item based on evidence, not intent. "We're planning to write a policy" is a 0, not a 1.
• Score across the organisation, not your best team. If engineering has great tooling but marketing is using free-tier ChatGPT with no guardrails, that's a 1 at best.
• Revisit quarterly. Readiness is not a one-time assessment. The AI tooling landscape shifts fast — Anthropic's amended S-1/A filing in November 2026 warned it "may imminently meet" its Responsible Scaling Policy threshold for AI-accelerated research, which signals capability jumps that will change what readiness looks like.

Where Most Organisations Fail: The People Measurement Gap

Of the seven checklist items, people skills baseline is the one most organisations skip or fake. They substitute tool adoption metrics ("80% of staff have a Copilot licence") for skill measurement ("here's what our people can actually do with AI").

The data backs this up. Across 2,482 AISA assessments, 45.6% of people scored in the Developing tier (composite score 28–59), and another 23.4% landed in Emerging (0–27). That means nearly 69% of assessed individuals are below the Proficient threshold. The average composite score is 46.3 out of 100.

The weakest dimension? Technical understanding, with an average score of 39.2 — meaning most people lack a working mental model of how AI tools actually function. This matters because without technical understanding, people can't diagnose why a tool gives bad output, can't select the right model for a task, and can't evaluate vendor claims.

The prediction gap compounds the problem. People overestimate their own AI skills by an average of 17.8 points. Students show the largest gap at 31.4 points; even engineering roles overestimate by 15.2 points. If you rely on self-assessment surveys to establish your people baseline, you're building on sand.

This is why checklist item #1 requires a named assessment tool and actual score distributions — not a survey asking people how confident they feel. Confidence and competence are different things, and the gap between them is where failed AI transformations live.

For a deeper look at what each skill level means in practice, see AI Fluency Levels: 5 Tiers Explained. And if you're wondering whether your team is keeping pace, Am I Falling Behind on AI? breaks down the data by role.

From Checklist to Formal AI Readiness Assessment

This checklist is a starting point — a structured way to identify gaps before you commit budget. But a checklist scored by internal stakeholders has inherent limitations: it relies on self-reporting, it doesn't produce individual-level data, and it can't benchmark you against external norms.

The natural next step is a formal AI readiness assessment that produces objective, evidence-based data. Here's what that progression looks like:

Step 1: Complete the Checklist

Run through the seven items with your cross-functional team. Identify which items scored 0 or 1. These are your priority gaps.

Step 2: Baseline Your People

For the people skills item specifically, move from "we don't know" to "we have data." An AI skills assessment gives you individual scores across defined dimensions, role-level benchmarks, and a clear picture of where training will have the most impact. AISA's assessment covers 11 criteria across five dimensions, produces evidence-linked scores (every rating is tied to what the candidate actually said), and delivers results that hold up on retake — 98% of retakers land in the same or adjacent tier.

Step 3: Address Policy and Governance Gaps

If items 2 and 4 scored low, these need to be resolved before scaling adoption. No amount of training helps if people don't know what they're allowed to do. The free Responsible AI Use guide covers the fundamentals of data privacy, acceptable use, and oversight design.

Step 4: Build Your Measurement Loop

Connect your people baseline to your organisational KPIs. Measure skill progression over time. Track whether training investments actually move scores. This is where most programmes stall — they run a one-off training session and never measure whether it worked.

Step 5: Iterate

Revisit the checklist quarterly. AI capabilities are shifting — new models, new agent frameworks, new regulatory requirements. Your readiness posture needs to keep pace. For a broader view of how to structure ongoing measurement, see AI Upskilling and Measurement in 2027.

If you've already completed a self-assessment and want to compare approaches, Is My Company AI Ready? walks through the self-assessment model in detail, and AI Maturity Assessment: How To covers the maturity-model approach.

---

Related reading: AI Fluency Levels: 5 Tiers Explained [2026] — what each score band means for individuals and teams.

Related reading: AI Upskilling and Measurement in 2027 [Strategy] — how to build a measurement loop that outlasts the pilot.

Related reading: Am I Falling Behind on AI? [2026 Data] — role-level benchmarks to see where your team stands.

---

Frequently Asked Questions

What is an AI readiness assessment?

An AI readiness assessment is a structured evaluation of an organisation's capacity to adopt, integrate, and sustain AI across its workforce and operations. It typically covers people skills, policy, tooling, governance, measurement, culture, and budget. The goal is to identify gaps before committing resources — so you invest in the right areas rather than buying tools your team can't use effectively.

How long does an AI readiness assessment take?

A checklist-based self-assessment like the one in this post can be completed in a single 90-minute cross-functional workshop. A formal assessment that includes individual skills measurement takes longer — typically one to two weeks to deploy across a team, depending on size. AISA's individual assessment takes approximately 20–30 minutes per person and produces results immediately, so the bottleneck is usually scheduling, not the assessment itself.

What's the difference between AI readiness and AI maturity?

AI readiness measures whether an organisation has the foundations to begin or scale AI adoption — skills, policy, governance, and infrastructure. AI maturity measures how far along that adoption journey an organisation has progressed, typically on a multi-stage model from experimental to optimised. Readiness is a prerequisite assessment; maturity is a progress assessment. You need readiness before you can meaningfully measure maturity. For a detailed walkthrough of the maturity-model approach, see AI Maturity Assessment: How To.