AI agent for product owners
Acceptance Criteria Test Mapping Agent
Make sure every accepted story has evidence for each of its acceptance criteria
What it does
At sprint review a story is marked done because it demos well, but half its acceptance criteria were never tested. At sprint end this agent reads each story's acceptance criteria and the linked test results, pull requests and demo notes. It maps every criterion to a test, a recorded demo or a note, and marks the ones with no evidence. It asks the team to add a test or a short demo recording, waits, and rechecks. It also checks that the linked test really covers the criterion and not just the same feature name. For each story it writes a short evidence summary. The product owner approves acceptance for each story. Edge case: a criterion says the page loads fast but no measurement exists, so the agent asks for a number.
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
- Sprint nearing its end
- Read each story and its acceptance criteria
- Collect linked tests, results, pull requests and demo notes
- Map each criterion to a piece of evidence
- Check that the evidence really covers the criterion
- List criteria with no or weak evidence by story
- Product owner approves the requests to the teamThe agent waits here for your OK.
- Send requests for tests or demo recordings and wait for replies
- Does every criterion now have passing evidence?If not: remap with the new evidence and ask again for what is still missing. Back to step 3.
- Product owner decides acceptance for each storyThe agent waits here for your OK.
- Evidence summary per story
How it decides
It counts a criterion as covered only when a passing test, recording or measured result points to it, and flags vague criteria that cannot be tested.
- Count a criterion as covered only with a passing test or recording
- Flag criteria that cannot be measured as vague
- Check that a linked test mentions the behavior, not just the feature
- Do not recommend acceptance when any criterion has no evidence
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Where evidence is stored
- Evidence types accepted
- Days before sprint end to run (default 2)
- How vague criteria are flagged
- Summary format
What keeps you in control
It always asks you first
- Product owner approves requests to the team
- Product owner approves acceptance of each story
Hard limits
- Never mark a story accepted
- Never edit acceptance criteria without the owner
It stops when
- Done: all stories have evidence or a decision on the gaps
- Stop: the sprint ends and uncovered stories move to the next sprint
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
