Insurance

Model risk does not stop at the model

Your model governance is thorough. Your code governance assumes a human wrote the code.

Insurers scrutinise the models they deploy. The code that implements pricing, underwriting and claims decisions gets far less attention, and it is now being written with the same AI assistance as everything else. PromptWake records that half of the picture.

Book a demo

Where it matters first

Rating engines, underwriting rules, claims triage and the reporting code that regulators read.
Book a demo

Implementation provenance, not just model provenance

Know which lines of your rating logic an agent produced, from which prompt, and whether a person revised them afterwards.

  • Line-level attribution
  • Prompt retained
  • Human revision separated

Evidence per period

Reports cover a date range, which fits the rhythm of internal audit and regulatory reporting rather than cutting across it.

A record that outlives the developer

Two years later, the person who wrote the prompt may have left. The prompt has not.

The question comes before the incident

Nobody asks how AI-written code entered a repository until something goes wrong or somebody audits it — and by then the record either exists or it does not. PromptWake is cheap to run before you need it and impossible to reconstruct afterwards.

Questions this sector asks

How does this relate to model risk management?

It covers the implementation layer that model documentation usually stops short of: the code that puts a model into production, and who or what wrote it.

Can we limit this to specific repositories?

Yes. Capture is per project, and a policy file can name the directories where AI authorship needs human sign-off.

Is the AI share a compliance metric?

It is a fact, not a target. A high figure is not a finding on its own — what matters is whether it was known, reviewed and recorded.

Underwriting code deserves the scrutiny the model gets.

Start with one rating repository and see what the record shows.
Book a demoSee the product

On AI accountability