Is my company AI ready? A scored checklist you can run today

Your company is AI ready when six things can each survive a specific test: your data, your processes, your people, your technology, your strategy and your governance. Not when you have a strategy deck. Not when someone has a ChatGPT licence. When each of the six passes a test you can run today, on your own, without a consultant in the room.

Below are the six tests, one per pillar, scored out of 20 each. Total 120. Run them in a single sitting and you will have a number, a weakest pillar, and a decision about whether to spend money on AI this quarter or spend it on the thing that will stop the AI working. This is the self-scored version of the free AI readiness audit, which asks 30 questions across the same six pillars in 7 minutes and hands you a score out of 120 with a band per pillar.

How to score it honestly

Each pillar gets a score from 0 to 20. Use this scale and resist the urge to round up:

• 0 to 5. You cannot answer the test question without asking someone else, and you are not sure who to ask.
• 6 to 10. You can answer it, and the answer is bad.
• 11 to 15. The answer is acceptable for most of the business, patchy in places you can name.
• 16 to 20. You could demonstrate it to a sceptical outsider inside an hour.

One rule. Score what is true this morning, not what will be true after the project that is currently at 40 per cent. Everybody who has ever failed at this scored the plan instead of the position.

Pillar one: data. The 500-customer test

The test: ask whoever owns your CRM for a clean, accurate list of your best 500 customers. Time how long it takes to arrive, and count how many manual steps sat between the request and the file.

Same day, no manual export-and-merge, and the definition of “best” is already agreed: 16 to 20. It arrives in two days after someone stitches three systems together in a spreadsheet: 6 to 10. Nobody can agree what “best” means and two people send you different lists: score it 5 and be grateful you found out now.

This is pillar one for a reason. According to Informatica’s CDO Insights 2025 survey of 600 global data leaders, 43 per cent cite data quality, completeness and readiness as the top obstacle preventing AI initiatives from reaching production. Gartner is blunter: through 2026, organisations will abandon 60 per cent of AI projects that lack AI-ready data.

The sequencing myth is the killer here. “We’ll fix the data once the AI is in” sounds pragmatic and is the most expensive sentence in the room, because the model amplifies whatever it is fed. I have written elsewhere about why dirty data is the cheapest AI mistake to fix first, and the cost asymmetry is the whole argument: cleaning a customer table is boring and cheap, discovering the model has been confidently wrong for six months is neither.

The audit asks five data questions, and the first is exactly the 500-customer one. The others cover completeness of decision-critical data, whether your systems can share data without a human moving spreadsheets between them, how often reliable data actually drives a decision, and who is responsible for data quality. If you cannot name a person for that last one, deduct five.

Score your data pillar: ___ / 20

Pillar two: process. The three-person test

The test: pick one core process. Ask three people in the same team to describe how they complete it, separately. Compare the steps.

Three near-identical descriptions: 16 to 20. Three descriptions that agree on the outcome but differ on the route: 11 to 15. Three genuinely different processes with different tools and different exceptions: you have not got a process, you have three habits, and you are nowhere near automating any of them.

This matters because AI does not automate an intention. It automates a defined sequence. If the sequence exists only in the head of the person who has done it since 2019, there is nothing to hand over. The audit’s process questions push on the same nerve: how many core processes are documented well enough that a new starter could follow them without asking anyone, whether three people would approach the same task the same way, how often you review processes rather than just running them, whether you know where the bottlenecks are, and who is accountable for process improvement.

A useful side effect: the three-person test finds your bottleneck for free. The step where the three descriptions diverge most is almost always the step where the work queues.

Score your process pillar: ___ / 20

Pillar three: people. The Tuesday test

The test: if you asked your team to use an AI tool in their work tomorrow, what happens? Not what they would say in a workshop. What happens on a normal Tuesday.

Most teams already use AI, quietly, on personal accounts, and will not tell you because they think it looks like cheating. That is a different score from a team that has not touched it. Both are different again from a team that has tried it, found it useless for their actual work, and stopped.

Score high if: your team’s attitude is curiosity rather than fear or performance, they would be capable tomorrow without a training week first, you personally use these tools and talk about it, and the business has put actual money into helping people understand them. Score low if your enthusiasm is the only enthusiasm in the building. Leaders consistently overscore this pillar. You are asking about a room you are in, and people modify their behaviour when you walk into it.

The capability question is not “can they write a prompt”. It is whether anyone in the team can tell when the output is wrong. That is the skill that decides whether AI saves time or quietly manufactures rework.

Score your people pillar: ___ / 20

Pillar four: technology. The integration test

The test: name your two most important business systems. Can they exchange data without a person exporting a CSV from one and importing it into the other?

If yes, and you can name how, score 16 to 20. If the answer is “there’s an integration but nobody’s sure it still runs”, score 11 or below. If the answer involves a named colleague and a monthly ritual, you have found the reason your AI pilot will stall at the demo.

A warning on this pillar. Three-quarters of CIOs surveyed by Dataiku reported remorse over at least one major AI vendor or platform selection made in the past 18 months. Buying more technology is the reflex when this pillar scores badly, and it is usually the wrong reflex. A low technology score is normally a plumbing problem, not a product gap. You do not need a new platform. You need the two you have to talk to each other.

Smaller businesses sometimes assume a low technology score disqualifies them entirely. It does not. There is a real path for a company doing AI without a tech team, it just starts somewhere narrower.

Score your technology pillar: ___ / 20

Pillar five: strategy. The finance director test

The test: state, in one sentence, the business outcome your first AI project will change, the number that will move, and when you expect to see it move.

Then say that sentence to your finance director and watch the face.

This is where most companies score worst while believing they score well, because they have a strategy document. A document is not the test. The test is one sentence with a metric and a date in it. “Improve customer experience” fails. “Cut average first-response time on support email from 9 hours to under 2, visible in the queue report by end of Q1” passes.

The pressure behind this is now measurable. A CloudZero survey of 260 finance executives, including 135 CFOs, found 87 per cent say they need to tie AI spend to business outcomes within the year, while only 22 per cent can do that today. Kyndryl’s 2025 Readiness Report found 61 per cent of 3,700 senior leaders feel more pressure to prove AI ROI than a year ago. And Teneo, in May 2026, found 65 per cent of CEOs report misalignment with their CFO on AI’s long-term value.

That misalignment is rarely about the technology. It is about payback windows. The CEO is describing a three-year position, the CFO is defending a twelve-month P&L, and neither has said so out loud. The one-sentence test forces the conversation early, when it is cheap.

Score your strategy pillar: ___ / 20

Pillar six: governance. The name test

The test: name the single person accountable for AI outcomes in your business. Out loud. Now.

If a name came instantly and that person would agree with you, 16 to 20. If you named a committee, 10. If you named a job title with nobody currently in it, 5. If you hesitated, 0, and that hesitation is worth more than the rest of this checklist put together.

Open Future Forum’s August 2026 finance data shows the vacuum widening rather than closing: business-unit sign-off on AI purchases fell from 18 per cent to 8 per cent inside a single month, while “no single owner yet” doubled to 14 per cent. The detail that should worry you: not one CEO or founder respondent reported an unowned AI purchase at their company, but one level down, 29 per cent of technology respondents did. The people at the top believe ownership exists. The people doing the work know it does not. I have gone further into who owns AI outcomes in a company and why a committee is not an answer.

Governance also covers the unglamorous half: what data your staff are allowed to put into which tools, who reviews output before it reaches a customer, and what happens when it is wrong. Score honestly. Most mid-sized businesses are at 5.

Score your governance pillar: ___ / 20

What your total means

90 to 120. You are ready. Pick one use case, define the metric, run it properly. Your risk now is scope, not readiness.

60 to 89. You are ready in parts. Do not launch across the business. Launch where your two strongest pillars overlap, and fund the fix on your weakest one in parallel.

30 to 59. Spending on AI now buys you an expensive proof that your data or your processes are not fit. RAND research published in 2025 found 80.3 per cent of enterprise AI projects fail to deliver promised business value, with 33.8 per cent abandoned before production and 28.4 per cent reaching production but missing expected value. Companies in this band are the ones supplying that statistic.

Under 30. The work in front of you is not AI work. It is data, documentation and accountability, and it will make the business better whether or not you ever deploy a model.

One more read of the numbers. Your total matters less than your lowest pillar, because AI projects fail at their weakest joint, not at their average. A company scoring 100 with governance at 4 is in more trouble than a company scoring 70 flat.

If you want the version that scores itself, the free audit asks 30 questions across the same six pillars, takes 7 minutes, and returns a score out of 120 with a band for each pillar and a short synopsis. Every answer is self-reported, so it is only as honest as you are. No sales call gets booked unless you ask for one.