Perplexity visibility services examine how your business and supporting sources appear during an agreed set of buyer research questions. The useful unit of review is not simply a linked domain: it is the claim, the source supporting it and the decision the reader is trying to make. TANTU AI checks factual context, source quality and the usefulness of the destination page, then develops a prioritized correction or content plan. The work helps your team understand the evidence behind an answer while keeping citations, recommendations and commercial outcomes distinct.
Look beyond whether the brand appears
A source can be linked without its brand being recommended. A brand can be mentioned through an independent publisher while its own website receives no link. Treating both outcomes as one score hides the work needed next.
We record the source page, the claim it supports, the brand’s role in the answer and whether the visitor can take a relevant next step. A supplier directory, product manual and customer story may each support a different decision.
Review accessibility before adding content
Perplexity describes PerplexityBot as a search crawler and recommends permitting it through the site’s crawler policy and hosting controls. The official Perplexity: Crawlers lists the current verification information. We review access under the site owner’s policy.
A robots rule alone cannot prove that a page was retrieved or selected. We also check whether the important information is available in readable HTML and whether the page can be reached through normal navigation.
Examine the evidence behind each gap
| Observed pattern | Question to investigate | Possible action |
|---|---|---|
| Competitor appears in a comparison | Does the source resolve a purchasing criterion we omit? | Add an evidence-backed comparison section. |
| Our product is described incorrectly | Which first-party or third-party fact conflicts? | Correct the source and record the change. |
| Our page is cited without buyer action | Does the page explain fit and the next step? | Improve the relevant inquiry or product route. |
| No usable answer returned | Was the platform or requested mode available? | Mark unavailable; do not score it as absence. |
These are diagnostic scenarios, not reported customer outcomes. A recommendation to create a page follows an evidence review; it is not an automatic response to every missing mention.
Use a stable, inspectable sample
We agree the market, language, account or access context, platform mode and question set. Repeated observations retain timestamps and source links where available. Changes in the collection method are logged so trends are interpreted correctly.
The report distinguishes your owned domains, third-party sources and competitor domains. It identifies the page or source requiring attention and assigns the follow-up to an owner. We do not describe a sampled set as the platform’s total audience or market share.
Scope source development responsibly
Where independent coverage would help a buyer verify a claim, we research appropriate publishers, partners and directories. Outreach is a documented attempt; editorial acceptance remains with the publisher. Paid distribution and publisher charges require separate approval.
Start with your priority market, the offer you can deliver and the evidence already available. Perplexity can be included as an agreed platform in a diagnostic or managed program; availability and access conditions are confirmed before collection.
Test the claim against the cited page
Read the sentence or answer segment associated with a citation, then inspect the source page. Ask whether it supports that specific claim, whether important conditions were omitted and whether the information is current. A link to your homepage may identify the business without supporting a detailed performance statement. Record a missing or ambiguous connection instead of treating every owned-domain citation as equally informative.
In an illustrative procurement answer, a cited product page may confirm a supported connection type while the summary implies compatibility with a whole operating environment. That difference matters to the buyer. The follow-up could be a clearer compatibility section, a correction to outdated documentation or no content change if the source already states the limit accurately. The record should preserve the answer, source, relevant claim and decision about what is within your control.
Trace repeated evidence to its origin
Several cited pages can repeat the same press release, product description or syndicated article. Classify source ownership and inspect where the claim originated before describing this as multiple independent confirmations. Distinct domains do not automatically establish distinct evidence. A useful source map connects the claim to its original support and notes later summaries or reproductions.
This method changes the next action. If the primary evidence is weak, additional distribution may repeat the weakness. If strong documentation exists but is difficult to navigate, improving the source page may be more useful. If genuinely independent evaluation would help the buyer, assess relevant publishers and permissioned evidence within the outreach scope. Acceptance means the source review is traceable and the resulting content brief resolves a defined information gap. It does not require a publisher to accept outreach or the platform to select a preferred source.
Verify the request type before interpreting access
Perplexity distinguishes its search crawler from user-triggered retrieval. Its documentation describes PerplexityBot as a search agent and Perplexity-User as a fetcher acting on a user request; these have different behavior around robots rules.
When investigating access, compare the documented agent and published IP information with actual server or firewall records. A user-agent string alone is not sufficient to verify a requester’s identity. The result informs a specific access fix, not a guarantee of a future citation.
Common questions
Do you sell Perplexity placements?
No. We provide research, content and measurement services. We cannot sell or guarantee a citation.
Can you measure all buyer conversations?
No. We inspect an agreed sample and available site data. Private user conversations are outside our access.
Sources & method
Platform facts are linked below. The planning steps, examples and acceptance criteria are TANTU AI’s operating approach. Examples are illustrative unless identified otherwise.
Reviewed 16 September 2026. Product availability and controls can change.