# Review results (/docs/guides/review-results)



Open **Findings**, then choose **Test results**. Select a run to see its customer-facing outcome and checks.

<DocImage src="/images/product/run-result.jpg" alt="Test result showing a needs-attention outcome with failed and passed checks" caption="The summary separates the overall outcome from each supporting check." />

<Tabs items="[&#x22;Summary&#x22;, &#x22;Activity&#x22;, &#x22;Finding&#x22;]">
  <Tab>
    The summary shows the outcome, attempt count, and every passed, failed, or
    inconclusive check. Start here when you need the decision.
  </Tab>

  <Tab>
    Activity shows the ordered actions, observations, verification steps,
    cleanup work, and material limitations. Use it to understand how the result
    happened.
  </Tab>

  <Tab>
    A repeatable problem may create or update a finding. The finding connects
    the expected behavior, observed behavior, customer impact, and supporting
    runs.
  </Tab>
</Tabs>

## Review in this order [#review-in-this-order]

1. Read the overall outcome and summary.
2. Review every failed or inconclusive check.
3. Open Activity and find the first observation that differs from the expected path.
4. Check whether the final product state was observed independently.
5. Read any cleanup or restoration result.
6. Open the related finding when the problem is repeatable.

## Interpret the outcome [#interpret-the-outcome]

* **Passed** means the expected state and protected behavior were verified.
* **Needs attention** means at least one meaningful check failed or a product problem was found.
* **Couldn't complete** means the run stopped, failed, or did not produce enough trustworthy evidence.

Do not count an inconclusive check as a pass. If the product state cannot be observed, fix the connection or verification path and run the journey again.

## Review a finding [#review-a-finding]

<DocImage src="/images/product/finding-detail.jpg" alt="Finding detail showing expected behavior, observed behavior, customer impact, and affected journey" caption="A finding turns run evidence into a product problem with a clear customer impact." />

Ask:

* Does the title describe the product problem rather than an internal error?
* Do expected and observed behavior conflict clearly?
* Is customer impact specific?
* Do the supporting runs show the same failure?
* Is the severity consistent with the demonstrated impact?
* Is there enough evidence to verify a later fix?

<Accordions>
  <Accordion title="What if the agent says it succeeded?">
    Use the verified product state as the result. The agent's own message is
    context, not proof.
  </Accordion>

  <Accordion title="What if evidence is missing?">
    Treat the relevant check as inconclusive. Fix the connection or verification
    problem before running the journey again.
  </Accordion>

  <Accordion title="What should I share with my team?">
    Share the finding and the minimum evidence needed to reproduce it. Keep
    credentials and unrelated customer data out of the report.
  </Accordion>
</Accordions>

Continue with [Fix and verify](/docs/guides/fix-and-verify).