The AI Readiness Audit
Twelve questions on whether your institution is adopting AI as infrastructure or as a collection of purchases.
A 12-statement self-assessment across 4 sections, scored out of 24. Around 6 minutes with a senior team. Score each statement No, In part or Yes, read the band, and use the lowest section as the place to start.
How to use this
A school is ready when it can state what each tool is for and stop using it without disruption. These twelve questions test that, across four layers.
Score each statement No (0), In part (1) or Yes (2). Answer for the institution as it behaves this term, not as it is described in policy. Where two people disagree, the lower score is the more useful one.
| No | In part | Yes | |
|---|---|---|---|
| 1We can name the specific teaching or administrative task each AI tool is meant to reduce. | |||
| 2Staff were shown what the tool cannot do, in plain terms, before it was introduced. | |||
| 3There is a written position on acceptable AI use that a teacher could apply without asking a leader. | |||
| 4Model outputs that touch a pupil record are reviewed and signed off by a named person. | |||
| 5We know which of our pupil data leaves the institution, to whom, and under what terms. | |||
| 6We could export our data and continue operating if a supplier failed tomorrow. | |||
| 7Training is scheduled and paid for, rather than absorbed into staff evenings. | |||
| 8Support is held by more than one person, so adoption survives a single departure. | |||
| 9Teachers can decline a tool without a professional cost, and that has actually happened. | |||
| 10We know the total annual cost of our AI estate, including licences, training and staff time. | |||
| 11We check whether AI use is narrowing or widening the gap between our highest and lowest attainers. | |||
| 12Renewal requires evidence of effect, not just an absence of complaints. |
Scoring by section
- Purpose and boundaries
- Statements 1, 2, 3. Out of 6. Your weakest layer is purpose. Tools without a stated task become their own justification, and the estate grows by accretion. Write down the task each tool is meant to reduce, in terms specific enough to be proved wrong within a term.
- Data and exit
- Statements 4, 5, 6. Out of 6. Your weakest layer is data and exit. An institution that cannot leave cannot negotiate, and cannot govern. Establish what an export actually contains before the next renewal, rather than after a failure.
- Capability in people
- Statements 7, 8, 9. Out of 6. Your weakest layer is capability. Adoption that lives in one enthusiastic member of staff is a single point of failure with a timetable. Fund the training, and spread the support.
- Evaluation and cost
- Statements 10, 11, 12. Out of 6. Your weakest layer is evaluation. Without cost recorded alongside effect, renewal becomes a habit rather than a decision. Take a baseline before the next adoption, not after it.
Reading the total
- 20–24 · Building
- You are treating AI as infrastructure: specified, evaluated and reversible. The exposure now is durability. Capability that depends on individuals is not yet institutional, and the next leadership change will test that.
- 13–19 · Adopting with gaps
- Real practice exists in some layers and nothing in others. That is the most common position and the least stable one, because the strong layers disguise the weak. Close the lowest-scoring layer before adding anything new.
- 7–12 · Purchasing, not adopting
- The estate is being bought rather than built. Tools arrive, staff absorb them, and nobody can say what changed. Start by writing down what each tool is for, so that something can be cancelled.
- 0–6 · Exposed
- On this assessment, decisions about teaching are being shaped by systems the institution cannot see into or leave. That is a governance position, not a technology one, and it will not be improved by another platform.
Take this evaluation online to see your score against the anonymised sector benchmark at educave.ai. educave.ai · Free to copy, forward and use inside your institution, with attribution.