What is an AI readiness audit, and what can't it tell you?

An AI readiness audit is a structured set of questions that tells you where your business currently stands across the six things AI depends on, before you spend money on AI. It is a diagnosis, not a treatment. The version on this site is free, runs to 30 questions across six pillars with five questions each, and takes 7 minutes. At the end you get a readiness score out of 120, a band for each pillar, and a short synopsis of what the pattern means.

The six pillars are Data, Process, People, Technology, Strategy and Governance, in that order. That order is not decorative. Data sits first because every pilot that falls over downstream of a model tends to have fallen over upstream in the data. Governance sits last because most businesses only think about it after something has gone wrong. A readiness score is a snapshot of whether the foundations underneath an AI project can carry the weight of one, expressed as a number and a band so you can take it into a board meeting without a slide deck.

What the audit actually asks

The questions are deliberately plain. They are the questions a sceptical operator would ask you over a coffee, not the ones a vendor would ask to qualify you.

The first Data question is: if you needed a clean accurate list of your best 500 customers right now, how confident are you in what you would get? Most people know the answer before they finish reading the sentence. The rest of that pillar asks how complete your decision-making data is, whether your key systems can share data without someone manually exporting and importing spreadsheets, how often reliable data actually drives an important decision rather than being retrofitted to justify one, and who is responsible for data quality by name.

Process works the same way. How many of your core processes are documented well enough that a new starter could follow them without asking anyone? If you asked three different people to complete the same task, how similar would their approach be? How often do you actively review and improve a process rather than just running it? Where are the bottlenecks, and who owns keeping the process honest over time?

People asks how you would honestly describe your team’s attitude to AI right now, how capable they would be if asked to use an AI tool tomorrow, whether you as the leader are modelling adoption yourself rather than delegating it, and whether the business has made any formal investment in helping the team understand it.

Technology, Strategy and Governance follow the same structure. Five questions each. No trick questions, no scoring tricks hidden behind the scenes.

What makes those questions useful is not sophistication. It is that most leadership teams have never answered all thirty of them in one sitting, and the pattern across thirty answers is more informative than any single answer. A business that scores well on Technology and badly on Data is not short of tools. It is short of something to point them at. We cover the data half of that in more detail in dirty data and the cheapest AI mistake to fix first.

What you get at the end

Three things. A readiness score out of 120. A band for each of the six pillars. A short synopsis.

That is the free tier in full. It does not hand you an action plan, and nobody books a call with you at the end unless you ask for one. The reason to be specific about this is that “free assessment” in this market usually means a lead capture form followed by a sales sequence, and the audience this is built for, CEOs, MDs, COOs, CFOs, CIOs and IT directors at businesses between 50 and 10,000 people, has been burned by that often enough to assume it by default.

If you want the action plan, there are paid tiers. The Full Report is £97 and gives you a much more robust report with action plans attached to the gaps. The Deep Dive is £497 and gives you a 30/60/90-day plan for fixing those gaps before you spend on AI. Both sit on the same diagnosis. You can take the free score and never buy either, and the score is still a real answer.

What the audit does not do

This is the part most assessment pages leave out, so here it is plainly.

It does not review the tools you have already bought. It will not tell you whether the Copilot licences you paid for in January are being used in October, or whether that pilot platform is still running. That is a separate exercise, usually a painful one, and it involves looking at usage logs rather than answering questions.

It does not audit your systems, data or code directly. Nothing connects to your CRM. Nothing scans your warehouse. No one reads your pipeline configs.

Every answer is self-reported. You answer the questions. The score reflects what you believe about your business, not what an external auditor has verified. That is a real limitation and it cuts both ways: an optimistic CEO will score higher than their business deserves, and a cautious CFO will score lower.

There is a way to use that, though. Have three people in the leadership team take it separately and compare the scores. Where the three of you disagree by two bands on the same pillar is more diagnostically useful than any individual number. If the CIO bands Data as weak and the CEO bands it as strong, you have found something before you spent anything on it.

It is not a call, a consultation or a demo. There is no calendar link at the end.

Why a 7-minute questionnaire is worth anything at all

The fair objection: thirty self-reported questions cannot possibly substitute for a proper assessment. Correct. It is not meant to. The question is whether it beats the alternative, which for most businesses is no diagnosis at all followed by a purchase.

The evidence that the alternative is not working is now fairly blunt. PwC’s 2026 CEO Survey found 56% of CEOs report neither increased revenue nor decreased costs from AI investments in the last 12 months, with only 12% reporting both (PwC 2026 CEO Survey). Infosys found 72% of senior executives at large US firms say fewer than 25% of their AI pilots reached full production while delivering intended ROI (Infosys via o-mega.ai).

Those are not technology failures. A model that works in a pilot and dies in production usually died because the data it needed at scale was not clean, or the process it was meant to automate was not the same process in three different departments, or nobody owned the outcome once the pilot team moved on. All three of those are visible in thirty questions answered honestly. None of them require a code audit to spot.

The governance numbers say the same thing from another angle. Deloitte found close to three-quarters of organisations plan to deploy AI agents within two years, while only 21% report a mature governance model for agentic AI (Deloitte 2026 State of AI). And only 35% of finance leaders believe their organisations effectively measure ROI from AI, despite 77% already using it in operations (FounderOperator). Deployment is running well ahead of the ability to govern or evidence it. If you are being asked to prove AI ROI to your board, a baseline score taken before the spend is the cheapest evidence you will ever produce.

Readiness is not maturity

These get used interchangeably in board papers and they should not be.

Readiness asks: can we start? Maturity asks: how far have we got? A business can be mature in the sense of having AI in production across three functions and still be unready in the sense that no single person can explain who signed off on the model that prices your quotes. Reporting maturity when you have not established readiness is how a board ends up confident about something it cannot defend.

The audit measures readiness. It is the first question, not the last one. Governance is a pillar inside it rather than a separate compliance exercise, which matters because the two are not the same thing either, as covered in AI governance vs compliance.

Who should take it, and when

Take it before you buy, not after. The score is most useful as a baseline, and a baseline taken six months into a programme tells you much less than one taken the week before procurement opens.

Take it if someone on your board has asked whether the business is ready and you do not currently have a defensible answer. Take it if you have run a pilot that performed fine and then quietly stopped. Take it if you have been asked to name the person accountable for AI outcomes and you had to think about it.

It is not for marketing leaders. The book serves them. The audit is built for the people who sign off the spend and carry the risk.

You can take the free AI readiness audit here. 30 questions, six pillars, 7 minutes, a score out of 120.

If you score badly, that is not a reason to delay. It is a list of what to fix first, in order, before anyone signs a licence agreement.