Home / AI Control Diagnostic / Sample report
See a sample report.
This is part of the report a two-week AI Control Diagnostic produces. The organisation is made up and every figure is sample data. Nothing here comes from a real client.
Executive summary
Where the organisation stands.
The first page of the report says it in plain words, for a board reader.
The sample organisation is a mid-sized insurer. It knows it has a gap in how AI is controlled. A policy exists, but nothing enforces it where people actually use AI.
It handles personal data and answers to a sector regulator, and it has already had an AI-related near miss. Put together, that gives a risk rating of Critical: exposure is well ahead of control.
The weakest areas are knowing what AI is in use and workforce capability. The strongest is governance and accountability. The report recommends containment first, then a plan over 90 days.
What the evidence showed.
Each of the six areas is scored from 1 to 5. Every finding is backed by interviews and documents, and the report says where evidence was thin.
Seeing what AI is in use 1.6 of 5
AI tools are tracked in informal lists that are known to be incomplete, and new use is found by accident. Fourteen unapproved tools turned up in network logs during the Diagnostic week.
Governance and accountability 2.2 of 5
An AI policy is published, but nothing enforces it at the point of use. The approval route exists on paper and is routinely bypassed.
Risk and regulation 2.0 of 5
Risk is discussed case by case, with no criteria and no audit trail. A claims pilot was processing full customer records without a data protection impact assessment.
Workforce capability 1.6 of 5
Everyone has the same AI access, whatever their skill, and training is a one-off awareness course. Appetite is high. The constraint is structure, not willingness.
Value and cost 1.8 of 5
The main licence costs are known and the rest of AI spend is not. Value is never measured once a tool goes live. Fewer than half the evidence items were provided, so this finding relies on what people said.
Adoption and reporting 2.0 of 5
Leaders get occasional updates when they ask. Managers have no view of AI use in their own teams, and staff are not sure what they are allowed to do.
Value
Where the value is.
The report lists the uses of AI worth controlling first, with an indicative yearly value for each. The figures are discounted for confidence, and none counts until a controlled pilot proves it.
Total: £1,296,832 a year, indicative.
Also in the report
The rest of the report.
The full sample runs to ten pages. We show it in an introduction, where we can explain how each part would apply to you.
- A baseline of workforce AI capability, and the recommended spread of Agent Passport® levels
- Ten actions over 30, 60 and 90 days, each with an owner
- A deployment plan configured for the organisation
- The evidence behind every finding, and its limits
Want this for your own organisation?
The AI Control Diagnostic takes two weeks and produces this report for you, with your own evidence. The price is published.