Verification
Did the change really happen, and did it help?
Example finding
Re-run the affected prompts
Mention and citation rates before and after a change are shown for the prompts it affects.
Quick answer: Verification
Verification checks that a change really landed on the target, then re-runs only the affected prompts to produce a before and after result. It is coming in V3.
Coming in V3. Arrives with the third release. This page describes the intended capability so you can see where the product is going. It is not available today.
What you see
Automatic checks that the target changed as expected, followed by a targeted re-run of only the affected prompts to produce a before and after result.
Target state check
The changed page is fetched and compared with the expected state.
Outcome status
Verified, failed or explicitly unverified.
Affected prompt re-run
Only the prompts tied to the action are re-run.
Before and after bundle
Observations from before and after, side by side.
How it works
- 1
Check
Confirm the observable state of the target after the change.
- 2
Re-run
Run the affected prompts instead of a full audit.
- 3
Compare
Report the difference with sample size and caveats.
How it is measured
Definitions come first, so numbers stay comparable from one audit to the next.
- Verification outcome
- Verified, failed or unverified.
- Affected scope
- The topics and prompts re-run.
- Delta
- Change in mention, recommendation or citation rate with sample size.
Good to know. Coming in V3. No action is marked successful without observable verification. Basic re-run planning is coming in V2.
Common questions
Does verification prove my change worked?
It shows what changed on a defined sample. It does not prove causation.
What if the check fails?
The action is marked failed, logged, and rollback is offered where the integration supports it.
Does it prove causation?
No. It shows what changed on a defined sample.
A short checklist
- Record the baseline
- Wait a sensible interval
- Check the target state
- Re-run affected prompts
- Write the conclusion, including null results
Who uses it
- Analyst
- A defined before and after method.
- Marketing lead
- Evidence that work landed.
- Agency
- A credible client story.
Common mistakes
Verifying too soon
Allow time for sources to be re-read.
Crediting every change to the action
Other factors move answers.
Changing the prompts
Use the same set.
See Verification on your own brand.
Start with a first audit. We show where you appear, who is named instead, and what to fix first.