There's a frustration specific to Perplexity that other AI engines don't produce as clearly.
You test your visibility. Perplexity knows your sector. It cites competitors. It cites generic articles. But not you — even though your content is more precise, more documented, more recent than what it chooses.
This isn't arbitrary. It's the result of a documented mechanism: a 2025 analysis of LLMs with web search showed that Perplexity visits an average of 10 relevant pages per query — but only cites 3 to 4 in its final response. Being in the "candidate set" isn't enough. You need to be in the 3 or 4 retained.
The difference between being visited and being cited comes down to precise signals. This article describes them.
How Perplexity's filter works
Perplexity operates in three successive layers:
Layer 1 — Retrieval: Perplexity queries the web and retrieves a set of candidate pages — typically about ten for a standard query.
Layer 2 — Reranking: Perplexity's "Sonar" algorithm re-ranks these pages according to its own signal hierarchy — different from Google or Bing ranking. This is where most sites lose their place.
Layer 3 — Synthesis and citation: Perplexity builds its response by extracting passages from the top 3 to 4 sources and cites them inline.
The problem with most companies: they optimise for layer 1 (being found) but ignore layer 2 (being retained by Sonar). These are different logics.
The 6 real reasons Perplexity excludes you
Cause 1 — Your content is "marketing fluff" to Sonar
This is the most common cause and the least openly acknowledged. Perplexity searches for "Answer Chunks" — autonomous text blocks that directly answer a question with verifiable facts. Content rich in marketing formulations ("recognised leader", "innovative approach", "tailored solutions") but poor in concrete data provides nothing Sonar can extract and cite.
Sonar's algorithm measures the "factual density" and "semantic clarity" of content — not its editorial quality in human terms. A little-known technical blog with precise data and dated sources will systematically be preferred over a well-written but vague corporate page.
Documented example: if your competitor publishes a price table with precise figures and you display "contact us for a quote", Perplexity will cite the competitor — every time. It cannot cite a mystery.
Cause 2 — You have no presence in earned media
This is the most discriminating factor on Perplexity, and the one that takes the most time to build. A Fullintel-UConn study presented in February 2026 analysed Perplexity citation patterns: 47% of citations come from journalistic sources, and more than 89% are earned media.
Sonar's logic is explicit: sources that appear in third-party publications are those it considers to have third-party verified authority. Your own site cannot self-validate.
This is where many B2B companies are structurally stuck: they've never invested in press or sector media, because it wasn't necessary for their Google SEO. On Perplexity, this absence is a dealbreaker.
Cause 3 — Your content isn't fresh enough
Perplexity applies a temporal decay factor that practitioners have documented as having a "sweet spot" of 30 days. Content not updated for 6 months progressively loses its position in Sonar reranking, even if its content remains relevant.
This signal differs from Google's, which values historical authority more. Perplexity favours freshness more aggressively — advantaging publishers that publish regularly and disadvantaging "showcase" sites that only publish occasionally.
Content cited by AI engines is on average 25.7% more recent than traditional organic results (Ahrefs, 2025).
Cause 4 — PerplexityBot is blocked
Perplexity states this in its documentation: sites that block PerplexityBot have significantly lower chances of being indexed and cited. Check your robots.txt for the mention PerplexityBot.
In 2026, blocking AI crawlers is considered a high risk: with 25% of traditional search traffic migrating to AI interfaces, a company blocking these crawlers creates a structural "Visibility Gap" in growing acquisition channels.
Cause 5 — Your entity isn't clearly defined for Sonar
Perplexity uses an entity recognition system — it identifies companies, people and concepts as distinct objects with attributes. If your company name, description and area of expertise aren't consistent across your different online presences, Sonar can't build a stable entity for you.
Symptoms: Perplexity describes you differently depending on how the question is phrased, or confuses you with another company. This is the sign your entity is "blurry" in its representation model.
Cause 6 — Your content is too conversion-oriented
Sales pages, landing pages and typical product pages are built to convince — not to inform. Perplexity looks for informational and factual sources. A page whose main objective is to push towards purchase provides little extractable material for an AI response.
This doesn't mean your product pages shouldn't exist — it means they can't be your only type of content if you want to be cited by Perplexity.
The Perplexity paradox: being good isn't enough
Here's what data reveals about Perplexity's citation mechanics, explaining why many companies remain frustrated despite good content:
Perplexity can visit your page, judge it relevant, and still not cite it — because another source scored slightly higher on a secondary signal. Citation isn't binary: there's no "found" vs "not found". There's "found, excluded at layer 2" — and this reality is invisible to most companies.
This is what researchers call the "attribution gap": the difference between sources Perplexity consults and those it finally cites. Being in the 10 visited pages without being in the 3-4 cited is the most common situation — and the most frustrating, because it gives the illusion the problem is indexing when it's actually in the reranking.
Fixes in order of priority
Priority 1 — Check technical access (immediate impact)
Verify that PerplexityBot is authorised in your robots.txt. Test your page loading speed — a high TTFB reduces your chances of being retained by Sonar. Submit your sitemap if not done.
Priority 2 — Transform your content into "Answer Chunks" (2-4 weeks)
Identify your 3 most important pages for your acquisition. Reformat them: direct answer in the first 2 sentences under each heading, quantified data with sources and dates, comparison tables in clean HTML, structured lists. Remove long introductions that delay the answer.
Priority 3 — Build systematic freshness (ongoing)
Set up a content update schedule for existing content — not just publishing new articles. Adding recent data, updating a statistic, or enriching an example is enough to reactivate the freshness signal. Aim for a significant update of your main pages each month.
Priority 4 — Build earned media signals (2-6 months)
This is the longest lever to build but the most durable. Identify the 5 publications regularly cited by Perplexity in your sector. Build a strategy to appear there: expertise contributions, responses to journalists, publication of studies others can cite.
What your competitors do differently
Companies dominating Perplexity citations in their sector generally share three characteristics not found among the invisible:
They publish original data. Studies, surveys, own data analyses — this type of content is cited by others, creating the earned media signals Sonar values.
They have a regular publication cadence. Not necessarily high — but regular. One publication per week for 6 months is worth more than 20 articles published in one month then silence.
They have consistent presence on third-party platforms. Active LinkedIn, presence in sector directories, mentions in comparison articles — as many entity signals Sonar can independently verify.
Where to start?
Start by identifying at which layer you're blocked. Our free scoring tool analyses your Perplexity visibility and the 4 other AI engines — with a structured score indicating whether you're in the "found but excluded" zone or "not found at all".
For a complete diagnostic with competitor analysis on Perplexity and precise identification of the causes of your invisibility, our AI Diagnostic delivers a structured report within 5 business days.
To go further, also read How to check if your business appears in Perplexity and Why ChatGPT ignores your business.
Sources: analysis of the attribution gap in LLMs (arXiv, 2025), Fullintel-UConn study on Perplexity citation patterns (IPRRC, February 2026), Ahrefs data on content freshness cited (2025), Perplexity official documentation on PerplexityBot (2025-2026).