Perplexity cited your page. What does that actually prove?

Imagine this: you fix an access problem on a page on Monday. On Wednesday, Perplexity cites it. It is tempting to call the fix a success, but those two events alone do not prove that the fix caused the citation.

That is a hypothetical example, but it gives you three separate questions to check:

1. Could Perplexity actually access the page?
2. Was the exact page cited?
3. Did the change cause the citation?

Here is a practical way to keep those findings separate.

1. Page access

Start with the request record: when it arrived, which URL it requested, the response status, and any firewall or security action. Check the returned content where you have it.

A 200 status alone does not confirm that the intended page text was served. Even a successful request that returned the right content does not establish that Perplexity used it in an answer.

It also matters which agent made the request. Perplexity's official documentation describes PerplexityBot as designed to surface and link websites in its search results. It recommends allowing this bot in robots.txt and permitting requests from its published IP ranges.

Perplexity-User serves a different purpose: it may fetch a page in response to a user's question and include a link in the answer. Perplexity says this fetcher generally ignores robots.txt because the fetch is user requested.

For request verification, Perplexity's firewall guidance combines the user-agent name with the source IP address. Use its current published ranges; the agent name alone is not enough.

Keep your finding as narrow as your evidence: what your server returned, to which request, and when. If you only have a status code, say so.

2. Exact-page citation

Open the citation. Follow any redirects and check the final URL against the page you are testing. A link to another page on your domain does not show that your edited page was cited.

Save the question, full answer, date, cited URL and the claim beside the citation. Then read the page: does it actually support that claim?

Also keep a source citation separate from a brand or product recommendation. Record what the answer says rather than treating every link as an endorsement.

3. Causation

Back to the Monday/Wednesday example: fixing access and later seeing a citation is an observation. Connecting the two as cause and effect takes more evidence.

Before crediting the repair, check what else differed. Was there another page update or a wider site change? Did you ask a different query? Were the competing sources or the generated answer different?

Those are things to investigate, not explanations you have already proved.

A careful conclusion would be: “Access was repaired, and a citation was later observed. We have not isolated whether that repair caused the citation.”

For a useful before-and-after record, keep:

• Page URL and the access issue, with evidence of the problem.
• What changed, when it changed, and how the fix was verified.
• Exact prompt/query, test settings and test date for each check.
• Full answer, cited URL and the claim beside the citation.
• Other changes that could affect the comparison.

Repeating comparable checks can show whether a citation recurs. Repetition alone still does not isolate its cause.

When someone says “our Perplexity SEO change worked,” ask:

• Did access improve?
• Did the exact page start getting cited?
• Or do we have evidence that the change itself caused the citation?

These are different claims and need different evidence. Your report should make clear which one you can support.

Disclosure: Posted by the Xtrusio team. This is a documentation-based audit workflow with hypothetical examples, not a customer case study or a claim of measured results.

References

• Perplexity. Perplexity Crawlers.