Brand Mentions in Perplexity: How Perplexity Decides Which Brands to Name

That makes a name and a link two different results. A review roundup can put your name in the answer while the citation goes to the roundup. Your blog can be cited for a general fact while your name never appears. Our guide to getting cited by Perplexity AI explains the retrieval layer behind both. This post stays on the names: where they come from, how shopping and Comet change who gets named, and how to read the gap between named and cited.

Where Brand Mentions in Perplexity Come From

Perplexity builds each answer from passages it retrieves, then cites them. Its own guide says you can "highlight answer text to check the sources referenced." That makes brand mentions in Perplexity traceable. Pick any sentence that names a brand, highlight it, and you see which pages put the name there.

Two other inputs can add names. Perplexity's memory can draw on a signed-in user's saved preferences and past searches, and it shows those references "in the sources." And a user can turn web search off for a thread, so the answer comes from the model alone. Both are exceptions. For a typical signed-out question, the names come from the cited pages.

Category questions lean on lists, marketplaces and reviews

Ahrefs' September 2026 snapshot counted every domain Perplexity cited across more than 3.1 million US queries on all topics. Reddit (21.6%) and YouTube (20.8%) led the top-50 share. Further down, two groups stand out for product questions:

• Retail and marketplace sites: Amazon 3.6%, Alibaba 2.0%, Target 1.3%, Home Depot 0.8%, Etsy 0.7% and Best Buy 0.7%, or 9.1% together.
• Review, testing and rating sites: Tripadvisor 1.5%, Consumer Reports 1.3%, TechRadar 1.1%, NerdWallet 1.1%, Yelp 0.9%, Trustpilot 0.8%, CNET 0.8%, PCMag 0.8%, RTINGS 0.6% and Tom's Guide 0.6%, or 9.5% together.

Each of those pages names several brands. When one is cited, every brand on it has a chance to be named.

Business software shows the same pattern. Analyze AI, a tracking vendor, classified 115,843 citations from 460 B2B prompts by page type in August 2026. For Perplexity, 34.4% went to lists, comparisons and reviews, 35.4% to brand websites and 25.2% to editorial pages. In Profound's data from August 2024 to June 2025, Gartner took 1.0% and G2 0.6% of all Perplexity citations.

Questions about one brand lean on that brand's pages

For questions about a single brand, the brand's own site and profiles are the obvious sources. Yext's October 2025 study of 6.8 million citations across Gemini, OpenAI and Perplexity found 86% came from sources brands manage or influence: websites (44%), listings (42%) and reviews. Those figures pool three engines and mix branded and unbranded questions, so read them as a direction, not a Perplexity rate.

If what Perplexity says about you by name is the problem, our brand presence audit traces each wrong sentence to its source.

The Name-and-Link Grid: Named but Not Cited, Cited but Not Named

We call this the Name-and-Link Grid. It treats brand mentions in Perplexity and citations of your site as two separate yes-or-no questions, so every answer lands in one of four cells. Each cell has its own cause and its own fix, which is why one "visibility score" hides more than it shows.

Why names outlast sources

The 2026 study "Don't Measure Once" repeated the same prompts within 24 hours. Perplexity's cited sources overlapped by a Jaccard score of 0.282, where 1 means identical. Its named brands overlapped by 0.492. The same brand often survives a change of source, because several cited pages name it. That's good news for brands that appear on many lists, and bad news for brands that rely on one.

How often each cell happens

On matched B2B prompts in the Analyze AI study, Perplexity named the tracked brand 41.7% of the time. The authors' explanation: when Perplexity cites a "top 10" article, all ten brands can reach the answer. Part of that rate is the named-not-linked cell at work.

The opposite cell is harder to size. As of October 2026, no published study we could check splits Perplexity's citations into named and unnamed. Semrush's June 2026 study found 62% of citations were linked without the brand being named, but it covered ChatGPT, AI Overviews, Gemini and AI Mode only. Don't borrow its number for Perplexity. Count your own cells instead, as our guide to tracking brand mentions in Perplexity shows.

How Shopping, Comet and Premium Sources Change Brand Exposure

Three features change brand mentions in Perplexity, either by adding new formats or by changing which brands get picked.

Product cards and Instant Buy

For shopping questions, Perplexity shows product cards inside the answer. Its Instant Buy help article says it ranks product listings with criteria similar to its sources, "based on authority and relevance," and that it "analyzes product reviews across multiple sites." It adds three points that matter to brands:

1. Data depth counts. "Merchants who provide deeper product details such as availability, reviews, pricing, and specifications are more likely to be recommended."
2. No paid slots. "Advertisers cannot pay to list their products," and listings are set "based on relevance and ratings."
3. Personal history counts. Perplexity uses "past searches, interactions, and saved preferences" to refine recommendations, so a signed-in shopper may see different brands from a stranger.

Instant Buy is available to all US users, per the same article. Its PayPal help page lists merchants such as Wayfair and "thousands of other merchants on BigCommerce, Shopware, and Wix platforms." Large retailers can also share product data through the free Merchant Program, which Perplexity says raises the chance of being a "recommended product."

Comet

Comet is Perplexity's Chromium-based browser. Its help center lists Perplexity search as its primary search engine, plus Ask and Summarize buttons for the page a user has open. The Comet page shows it on Mac, Windows, iOS and Android, with shopping among its assistant's jobs.

Two effects follow. More of a Comet user's searches run through Perplexity, so the same sourcing patterns apply to more of their questions. And when the assistant summarizes your own page, your brand is the page, so your on-page facts carry straight into the answer. Perplexity doesn't document how Comet's assistant chooses between merchants beyond the product-ranking rules above.

Publishers have a separate stake. Comet Plus pays partner publishers on "human visits, search citations, and agent actions." It pays for being cited. It isn't described as changing which brands get named.

Premium data sources

Perplexity lets users pick premium sources such as PitchBook, which offers "detailed firmographics for private and public companies," along with CB Insights and Statista. When a user picks one, company facts may come from that database instead of the open web. Keep your profiles there accurate if investors or analysts research you.

Worked Example: Who Names Tallyfold?

Tallyfold is a fictional invoicing and payments app for agencies. Its rivals, Brindlework and Kestrelyn, are fictional too, and every number below is illustrative.

To map its brand mentions in Perplexity, Tallyfold runs 10 prompts three times each in incognito mode: five "best X" prompts, three comparisons and two questions about Tallyfold itself. That's 30 answers. It sorts each one into the grid:

Tallyfold's mention rate is (5 + 7) ÷ 30 = 40%. Its citation rate is (5 + 4) ÷ 30 = 30%. The two numbers tell different stories, which is the point of keeping them apart.

Highlighting the seven named-not-linked sentences shows where the names came from: three from one review site's agency invoicing roundup, two from a Reddit thread, one from an app marketplace listing and one from a YouTube walkthrough (3 + 2 + 1 + 1 = 7). Tallyfold has no page of its own that answers "best invoicing software for agencies," so the roundup speaks for it.

The four linked-not-named answers all cite Tallyfold's guide to late-payment fees, used for a general fact in category answers. The guide never says how Tallyfold handles fees. The fixes follow the grid:

1. Named, not linked: publish a fair comparison of agency invoicing tools, and send the review site Tallyfold's current features and prices.
2. Linked, not named: add a short, factual section to the fees guide on how Tallyfold's reminder settings apply the method.
3. Neither: check the prompts with no Tallyfold presence against the lists Perplexity cites, and pitch the ones that fit.

What Moves Brand Mentions in Perplexity

• Be on the lists Perplexity cites. Find them by highlighting named sentences in your category's answers.
• Publish your own answer page for each "best", "vs" and "alternatives" question buyers ask, with a dated comparison table.
• Keep product and company data complete wherever Perplexity reads it: product feeds, marketplaces and data providers.
• Join community threads openly, as staff, with facts and links.
• Let PerplexityBot in. A page it can't fetch can't name you. Our PerplexityBot user agent guide covers access.

For the optimization side across both engines, see how to optimize brand mentions in ChatGPT and Perplexity.

Where Rankbox Fits

Rankbox doesn't track brand mentions in Perplexity or any other engine, so it can't tell you which cell you're in. The free AI Visibility Prompt Kit gives you 30 buyer prompts to run by hand, with a scorecard. Where Rankbox helps is the fix list. Answer-Space Research maps the questions buyers ask Perplexity, and the Citation-Ready Writer writes source-backed comparison and answer pages for the questions where a roundup speaks for you. The Business plan is $49.50 a month with a 7-day trial: see pricing.