AI readiness assessment

An AI readiness assessment is a structured check, done before money is committed, of whether a company can make AI work. The term covers very different things, and it is worth knowing which one you are being offered.

Questionnaires and maturity models ask people how things are: do you have data owners, is there a data dictionary, how often are records audited. They are quick and cheap, and they measure what people believe. They can’t see the customer who is active in the CRM and missing from billing.

Vendor readiness checks are run by the firm selling the AI, usually on its own system’s data. They are often useful technically, but the firm doing the checking is paid when you sign.

Evidence-based assessments work from the company’s own records, compared across systems, and test them against the AI being proposed. They take longer than a questionnaire and say more, because they find the gaps nobody knew to report.

Whatever the form, ask four things of an assessment before you rely on it: what was measured and what wasn’t; whether records were compared across systems or only within one; whether the assessor sells the AI, the software or the fix; and whether the output is a verdict you can act on, or a maturity level.

Our version is the Truth Score Read: exports from two or more systems, compared record by record, with a signed memo carrying the Truth Score™, a Leverability read on the AI in question, the cross-system red flags and a verdict. For every AI bet at once, there is the AI Investment Review.