AI readiness audit: what it is and how to run one

An AI readiness audit is a structured review of whether your organisation can turn AI activity into a measurable return. It looks at five things: governance, ownership, alignment, capability and control. In plain terms, it tells you whether the AI work already happening in your business is set up to produce value you can point to, or whether it is spending time and budget without a clear line back to a result.

Most UK businesses are past the question of whether to use AI. The harder question is whether the pilots, tools and policies now in place actually add up to anything. An audit answers that. It gives you an honest read on where you stand before you commit more money to the next round of tools or hires.

If you want a fast starting point, you can run the free benchmark and get a personalised report showing how your position compares with peers.

What is an AI readiness audit

A readiness audit is not a technology review and it is not a list of tools you could buy. It is an assessment of your organisation’s ability to deliver returns from AI and to keep doing so as the work scales.

The distinction matters because readiness gaps are rarely about the technology. A business can have capable models and useful software in place and still get nothing back, because no one owns the outcome, the policy sits unread, or the person leading the work has a mandate but no team to deliver it. An audit surfaces those gaps in a way that a tool demo never will.

This is worth doing if AI has started in your business but has not reached production, if you have written a policy that is not applied consistently, or if you have a funded AI owner with budget and mandate but no team behind them.

What an AI readiness audit looks at

A useful audit covers five areas.

Governance. Do you have a policy for how AI is used, and is it actually followed day to day? A policy that exists only as a document does not count as governance.

Ownership. Is there a single named person accountable for AI outcomes, or is activity spread across teams with no one answerable for the result?

Alignment. Is the AI work tied to a business objective that leadership cares about, or is it a set of experiments running to their own logic?

Capability. Do the people involved have the skills, time and support to deliver? A funded lead with no team is a capability gap, not a resourcing detail.

Control. Can you see what AI is doing across the business, measure it, and step in when something goes wrong? Without control you cannot report on return, and you cannot manage risk.

Each area is scored on evidence, not intention. The question is always what is happening in practice, not what the plan says should happen.

How to run an AI readiness audit step by step

You can run a first pass in a week. A fuller version takes longer but follows the same shape.

1. Set the scope. Decide whether you are auditing the whole organisation or one function. A single function is a sensible place to start.
2. Map current AI activity. List every AI tool, pilot and workflow in use, including the informal ones people have adopted without sign-off. You will usually find more than you expected.
3. Assign the five areas. For each area (governance, ownership, alignment, capability, control), gather evidence and score where you actually are.
4. Find the gaps. Look for the places where activity is happening but one of the five foundations is missing. That is where value is leaking.
5. Prioritise. Rank the gaps by how much they hold back return. A missing owner usually outranks a missing tool.
6. Write the actions. For each priority gap, name the fix, the owner and the date. An audit with no owned actions changes nothing.
7. Set a review date. Book the next audit before you close this one.

Who should own the audit

The audit needs a single owner who is senior enough to see across functions and to act on what they find. In practice this is often the person already accountable for AI, or a member of the leadership team if no such role exists yet.

It should not be delegated to whoever happens to run the most pilots. The point of the audit is to judge the whole picture, including whether ownership itself is clear, and that is hard to do from inside one project.

Common findings

Across UK businesses the same patterns come up. A policy has been written but is not applied, so governance exists on paper only. AI activity is under way but has no single owner, so no one can be held to the outcome. A business has funded an AI lead with real budget and mandate, but has given them no team to deliver, so the mandate stalls.

None of these are failures of technology. They are gaps in how the work is set up, which is exactly what an audit is built to catch.

For the fuller picture on what readiness looks like, see our related guide to the AI readiness assessment.

How often to repeat it

Run an audit at least twice a year, and again after any significant change: a new AI owner, a move from pilot to production, a new tool rolled out across teams, or a shift in your objectives. AI moves quickly enough that a picture six months old is often out of date.

The pattern that works is a light quarterly check against your last set of actions, with a fuller audit once or twice a year.

Our findings on readiness draw on the State of UK AI Adoption Survey 2026, 755 UK business leaders, in partnership with Executive Summary, which you can read in our research.

Run your first-pass audit

The fastest way to run a first-pass audit is to Benchmark your AI ROI. It takes a few minutes, shows you how you compare with peers, and gives you a personalised report you can use as the starting point for the full audit above.