The AI readiness checklist: 12 questions to score your business

TLDR:

• An AI readiness check scores one business process at a time, not the whole company.
• This checklist uses 12 questions in four areas: problem clarity, data, process stability and team readiness.
• Each question scores 0–3, for a maximum of 36 points.
• 25 or more means the process is ready to build on; 13–24 means fix the weak areas as part of the project; 12 or less means do the groundwork first.
• The lowest-scoring area tells you what to work on next.

An AI readiness checklist scores whether a specific business process is ready for AI before any money is spent on building. This checklist is the self-assessment version of the first stage of a ForceMX audit. It takes about 30 minutes per process and works best when the person who does the work, their manager and someone who knows your systems each score it separately.

Score one process at a time, for example "writing sales proposals" or "processing supplier invoices". Scoring "the company" gives an average that hides the useful detail.

How do you score problem clarity?

Problem clarity measures whether you can say what the process costs today and who it affects. Without that, nobody can show later that the project paid off.

1. Can you describe the problem in money or hours?

• 0: "We want to be smarter about this." No number attached.
• 1: "It takes a lot of time." Still no number.
• 2: "It costs roughly 60 hours a month." Estimated, not measured.
• 3: "It costs 60 hours a month; we measured it from time records or a two-week log."

2. How many people or teams does the problem touch?

• 0: One person, occasionally.
• 1: A few people in one team.
• 2: A whole team or department.
• 3: Several departments, or customers directly.

3. Is the work done by hand today?

• 0: Existing software already does it well.
• 1: Software does most of it and people fix the rest.
• 2: Software exists but people work around it.
• 3: It is almost entirely manual: copying, checking, writing, chasing.

How do you score data readiness?

Data readiness measures whether the information the process needs exists, is consistent and can be reached. AI output is only as reliable as the data it works from.

4. Is there a record of past cases?

• 0: No record; each case is handled from scratch.
• 1: Scattered across email, chat and personal spreadsheets.
• 2: In one system, but with duplicates and gaps.
• 3: In one system, reasonably complete, with at least several months of history.

5. Is the data recorded consistently?

• 0: Every team records it differently.
• 1: There is a standard, but few follow it.
• 2: Most records follow the standard.
• 3: The system enforces the standard.

6. Can the data be reached without a long IT request?

• 0: Locked in an old system that nobody can export from.
• 1: Exports are possible but slow and manual.
• 2: Available through reports or an API with some technical help.
• 3: Available through an API or integration the team already uses.

How do you score process stability?

Process stability measures whether the work follows rules that can be written down and that stay put. A process that changes every week will break any automation built on it.

7. Does the process follow consistent rules?

• 0: It is handled differently every time.
• 1: There are guidelines, with many exceptions.
• 2: The core steps are stable, with occasional changes.
• 3: It is documented and changes are planned.

8. Is success defined and tracked?

• 0: Nobody has defined what good looks like.
• 1: Goals exist but are vague ("faster", "fewer mistakes").
• 2: There are metrics, tracked now and then.
• 3: There are clear metrics, tracked at least monthly.

9. Do people agree the process needs fixing?

• 0: Management wants it; the team sees no problem.
• 1: Some agree, some do not.
• 2: Most agree there is a problem, not on how big.
• 3: Broad agreement that it costs real time or money.

How do you score team readiness?

Team readiness measures whether the business can adopt a new way of working and keep it running. A system nobody uses or maintains delivers nothing.

10. How have past changes to tools or processes gone?

• 0: New tools rarely stick.
• 1: Some stick, many fade.
• 2: Most stick within a few months.
• 3: New tools are usually in daily use within weeks.

11. Is there an owner and a budget?

• 0: Interest, but no owner or budget.
• 1: Budget exists but competes with other priorities.
• 2: An owner is named and budget is set aside.
• 3: Owner, budget and a place on the roadmap.

12. Who will build and maintain it?

• 0: Nobody identified.
• 1: One person who might have time.
• 2: A technical team without AI experience.
• 3: In-house capability or a partner who will support it after launch.

What does your score mean?

Add the 12 scores for a total out of 36, then read the band and the weakest area together.

The band matters less than the lowest-scoring area. A process can score 28 overall and still fail if data readiness scores 2 out of 9.

What should you fix first?

Fix the lowest-scoring area first, because it is the one most likely to stop the project. Common patterns and their fixes:

• Data is the weak area: agree one place and one format for the records, and start capturing them properly. This often takes a few weeks and pays off even without AI.
• Process stability is the weak area: write the process down as it is really done, agree the rules, and remove the exceptions nobody needs.
• Agreement is the weak area: run a small pilot with the most willing team and share the measured result.
• Ownership is the weak area: name one owner with the authority to change how the work is done.

For the reasons these gaps matter, see why most AI projects fail.