A shipped change is not a result
The same prompt, re-probed after the work ships. Sonarvue compares the answer, the citations, and the competitor set, then keeps the delta and the audit trail attached to the action.
Compare the same prompt
AI answers move for reasons that have nothing to do with your work: a model update, a new source, a competitor's launch. A before-and-after screenshot from two different questions proves nothing. Verification starts with the exact prompt, engine, market, and language from the baseline, then runs it again after the change.
You see the answer side by side: whether you were named, where you appeared, how you were described, which sources were cited, and which competitors took the shortlist. The comparison is stored with the action that preceded it.
What you see
- Answer delta. The sentence that changed, not just a score.
- Citation delta. The sources added, removed, or replaced.
- Competitor delta. Who entered the answer, who left, and who moved ahead.
- Action record. What shipped, who approved it, when it went live, and what evidence supported it.
When the answer does not move
A flat result is useful. The agent records the null result, checks whether the page was indexed and cited, and routes the next diagnosis from what the engine actually did. You can stop spending on a fix that did not matter, or narrow the gap and try the next source.
The audit trail
Every step keeps its evidence: the baseline answer, the diagnosis, the approved action, the publish record, and the re-probe. That gives an agency a client-ready record and gives an in-house team a defensible history of what changed and why.
Verification FAQ
How is this different from rank tracking?
Rank tracking measures a page's position in a list of links. Verification measures what the answer engine said after the work shipped: whether you were named, cited, and described correctly. A page can rank and still be absent from the AI answer.
How long after a change should we re-probe?
It depends on the engine and the type of change. Sonarvue schedules the first check after the page is live and indexed, then keeps checking on the cadence you set so a one-off answer does not become a false positive.
What if the answer changes on its own?
The baseline, the date, and the run history make drift visible. The agent separates a change that followed your action from movement that happened across the category at the same time.
Can I show this to a client?
Yes. The evidence trail includes the prompt, the answer, the action, the approval, and the re-probe. It is designed to be handed over without an explanatory call.
Prove the next change moved the answer. Start with a baseline, then keep the evidence attached to every action.
Find your first gapRelated: technical SEO · authority · how it works