AI agent for auditors
Change Management Sample Testing Agent
A consistent, documented sample test of change management, with the extent of any weakness known
What it does
Sample testing of changes is slow and inconsistent. This agent draws a sample of changes from the period under review and checks each one for approval, testing and deployment matching the ticket. It records exceptions with the reason. If exceptions are above a set threshold, it pulls extra samples to see whether the problem is isolated or widespread. It then drafts a result summary for the audit file. The auditor approves the conclusions. Edge case: an emergency change was approved after deployment, so the agent checks it against the emergency change rule before calling it an exception.
How it works
Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.
Read the steps as a list
- Audit test is due
- Pull the population of changes for the period
- Draw the sample using the sampling rules
- Check approval, testing and deployment evidence for each change
- Record exceptions with the reason
- Are exceptions below the threshold?If not: draw additional samples and test them the same way. Back to step 2.
- Draft the result summary and conclusion
- Does each conclusion cite evidence in the sample?If not: add the evidence or remove the claim. Back to step 7.
- Auditor approves the conclusionsThe agent waits here for your OK.
- Test workpaper
How it decides
A change passes when approval came before deployment, testing is recorded and the deployment matches the ticket. Exceptions above the threshold widen the sample.
- Count approval after deployment as an exception unless it is an emergency change
- Widen the sample when exceptions exceed the threshold
- Check deployment logs against the ticket
- Cite evidence for every conclusion
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Sample size
- Exception threshold (default: 10%)
- Period under review
- Emergency change rule
What keeps you in control
It always asks you first
- Auditor approves the conclusions
Hard limits
- Never changes records
- Never states a conclusion without evidence
It stops when
- Done: the test is complete and approved
- Stop: the population cannot be confirmed as complete
Set it up
We guide you through the set-up, step by step
Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.
- One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
- The agent then walks you through connecting your own data, one source at a time
- A downloadable copy with the flow chart, the rules and the full guide