AI Adoption · Written and maintained by Haink’s AI adoption team · Updated July 2026 · 6 min read
How to Assess AI Readiness: A Step-by-Step Method
To assess AI readiness, gather the right people, score each of five dimensions on real evidence, read the shape to find your weakest link, and turn that into a first move — a pilot if you’re ready, a prerequisite to fix if you’re not. The method is simple. The discipline it demands — scoring what is true rather than what you hope is true — is the hard part, and the part that makes the result worth anything.
This is the running method. For the underlying model — what each dimension means and how the scoring works — see the AI readiness assessment framework; for the concept itself, what AI readiness is.
Step 1 — Set the scope
Decide what you are assessing: the whole company, or a single business unit. For a first assessment, a unit is often the better target — it is concrete, the people who know the truth are reachable, and the result maps directly to a first use case. A company-wide assessment is valuable later, when you are sequencing a broad program.
Step 2 — Get the right people in the room
Each dimension should be judged by the person who owns it. A readiness assessment scored by one function inherits that function’s blind spots — IT will overrate the organizational dimensions, the business will overrate the technical ones. Bring:
- Business leadership — for strategy and value.
- The data leader — for the honest state of the data.
- Process owners — for the workflows AI would touch.
- The technology leader — for execution and infrastructure.
- Risk & compliance — for the constraints that shape what can be built.
Step 3 — Gather evidence, not opinions
The single thing that separates a useful assessment from a comforting one is evidence. For each dimension, look at something real; score the example that exists, not the plan that doesn’t.
| Dimension | Ask to see… |
|---|---|
| Strategy | A written business goal AI would serve, and the name of the executive who owns it. |
| Data | An actual dataset a candidate use case needs — where it lives, who can access it, how clean it is. |
| People | The specific skills on hand and the named champion who will carry delivery. |
| Processes | A mapped workflow with a metric — evidence the process is understood well enough to improve. |
| Execution | A software project the organization actually shipped and now operates. |
If the evidence for a dimension doesn’t exist, that is the finding — and usually a low score.
Step 4 — Score each dimension
Rate each dimension on a simple scale — a 1–5 is enough — anchored to the evidence you just reviewed. Write a one-line justification next to each score; the justification is what makes the number defensible and the conversation honest. (The free score does this conversion to a 0–100 result automatically.)
Step 5 — Read the shape and find the floor
Resist reading only the total. The weakest dimension usually decides the outcome, because AI initiatives break at the weakest link, not the average. A profile that is a balanced three everywhere is more ready than one with two fives and a one. Find the floor, and ask a single question of it: is this dimension a prerequisite for the use cases you have in mind?
Step 6 — Turn the result into a first move
An assessment that doesn’t change what you do next was theater. Two clean outcomes:
- The floor is high enough → move to choosing a first use case; see use-case prioritization.
- A critical dimension is low → fix that prerequisite before piloting. If it’s data — the usual culprit — start with data readiness for AI, then reassess.
Common mistakes when running the assessment
- Scoring aspirations. “We’re building that” is not evidence. Score today.
- One function scoring alone. Blind spots become the result. Get the owners of each dimension in the room.
- Reading the total, not the shape. A high average can hide a fatal low.
- Doing it once. Readiness moves; reassess after you close a gap, and before each major new initiative.
Run it in three minutes. The free AI Readiness Score walks this method for you — 11 questions across the five dimensions, an instant 0–100 result and a fix-first plan, no email required. When the decision needs an evidence-graded, board-defensible answer, the AI Adoption Assessment facilitates it and grades a deeper profile — the difference is laid out in score vs expert assessment.
The honest verdict: the goal is to find the weak spot, not to pass
A readiness assessment is not an exam to score well on. Its entire value is in surfacing the one dimension that would sink your first project — while it’s still cheap to fix. Teams that go in hoping for a high number learn nothing; teams that go in hunting for their weakest link leave with a plan. Assess honestly, find the floor, fix it, and the pilot takes care of itself.
Frequently asked questions
How do you assess AI readiness?
Set the scope, get the owners of strategy, data, processes, delivery and risk in one room, score five dimensions on real evidence, read the shape to find the weakest dimension, and turn that into a first move.
Who should be involved?
The owners of each dimension — business, data, process, technology and risk/compliance. One function scoring alone produces that function’s blind spots.
What evidence do you need?
Concrete examples per dimension: a written goal and sponsor, an accessible dataset, real skills and a champion, a mapped process, a shipped project. Score what exists, not what’s planned.
How long does it take?
A self-assessment takes minutes with a structured score, or one workshop for a leadership team; an expert, evidence-graded assessment runs from about three weeks.
Can you assess it yourself?
Yes, for a directional baseline — a free score does it in minutes. The limit is that self-assessment measures what you can see; for a budgeted decision, an expert grades on evidence.
Assess your readiness now
The free AI Readiness Score runs this method in three minutes — an instant 0–100 result across five dimensions and a plan for your weakest one. No email required.