What is an AI readiness audit, and what should one actually produce?

A founder I spoke to last month had been quoted forty-eight thousand dollars for an AI readiness audit. The deliverable was a slide deck and a recommendation. The recommendation was a twelve-month implementation engagement. The audit was a sales mechanism, and a fairly expensive one. This piece is what an AI readiness audit actually is, what it should produce, and how to tell the real thing from the dressed-up version.

What an AI readiness audit actually is

An AI readiness audit is a structured, scored read of how prepared a business is to adopt AI in production. The output is a snapshot in time across a fixed set of dimensions, scored against a rubric, with a clear list of moves to make next.

It is not a strategy. It is not a transformation roadmap. It is not a vendor recommendation. It is the equivalent of a structural inspection on a house: a written read on the bones, the wiring, the foundation, before anyone agrees to a renovation.

A useful audit ends in three things, in order:

1. A score. A headline number against a fixed scale, so the read is comparable to itself six months later and across businesses.
2. A prioritised action list. The three to five moves with the highest expected return per week of effort, plus two or three things to actively stop doing.
3. An honest go or stop. Sometimes the answer is “not yet, fix these two upstream things first.” Audits that always recommend AI adoption are not audits.

The five dimensions worth scoring

Most readiness frameworks list eight, ten, or twelve dimensions. Most of those collapse on inspection. The five that earn their keep:

1. Data. Is the data the AI would use accessible, clean enough, and recent enough to support the use case? The failure mode is rarely “no data”; it is “data scattered across five tools, none of them speaking to each other.”
2. Process. Are the workflows the AI would touch documented enough that automation has something concrete to bite on? AI accelerates clear processes and accelerates the failure of unclear ones equally.
3. People. Does anyone in the building actually own the AI work, or is the assumption that the vendor will? Adoption without an internal owner stalls at month four every time.
4. Governance. Is there a written AI use policy, a vendor diligence position, and a clear data classification scheme? Governance is the cheapest thing to fix and the most commonly missing.
5. Economics. Is there a real, sized value at stake worth the investment? Or is this a fashion purchase because the board asked “what is our AI strategy” in the last meeting?

A score per dimension, an honest readout, a prioritised list of moves: that is the audit. Anything else is decoration.

What good output looks like

A useful AI readiness audit produces, in order:

• A one-page summary with the headline score, the five dimension scores, and the top three moves.
• A scored read on each dimension with the specific evidence that drove the score, in plain English.
• A prioritised action list across the next twelve weeks, with estimated effort and expected return per move.
• A short list of things to defer or stop, named specifically.
• A go or stop on AI adoption right now, with the rationale.
• A re-audit date and what to measure between now and then.

The deliverable is yours to keep. It is not a teaser for a follow-up engagement; it is the engagement.

How to tell a real audit from a dressed-up discovery call

The differences are visible on the proposal page, not in the delivery. Patterns that flag the discovery call masquerading as an audit:

• The price is quoted as a percentage of the implementation project, not a fixed fee.
• The deliverable is a presentation, not a report.
• The score is qualitative (“medium-high readiness”) with no rubric you can see.
• The next-steps section is longer than the assessment section.
• The audit and the recommended follow-up work are sold by the same firm with no third-party option in the report.

None of these is dishonest. They are, however, a different product. If the conversation is moving toward a six-figure implementation, run the audit somewhere else first. The cost of a second opinion is trivial.

What an AI readiness audit costs

For a small to mid-sized business, a properly scoped audit should sit between one hundred and one thousand dollars. Free taster scores are useful for the headline number; the full audit earns its keep on the prioritised action list and the deliverable you keep.

The DPEX AI Readiness Audit ships in two tiers: a free two-minute Mini-Score that gives you the headline number, and a $195 AUD full audit that produces a seventeen-page branded PDF with the scored read, prioritised actions, and a twelve-week plan. Both are scored against the same fixed rubric, so the score reads the same six months later when you run it again.

The honest summary

The AI readiness audit is one of the highest-leverage, lowest-cost moves a business can make before committing real money to AI adoption. It is also one of the most over-priced deliverables on the market, because firms have learned to charge for it as if it were the implementation. Pay for the audit, not the sales pitch. Keep the deliverable. Re-run it every six months. That is the practice.

The Mini-Score below is free; the full audit ships in twenty minutes of your time and one hundred and ninety-five dollars. Both are honest reads, scored the same way.