How to Get Cited by Perplexity: 9 Source Signals

We put 500 Houston real estate questions to four AI surfaces in September 2026. Perplexity named more firms than any of the others, yet 12.0% of its answers named no firm at all, and an estimated 94.0% of the Houston Association of Realtors’ 50,000-plus members were never named in 2,000 answers. That gap is the opportunity. Our figures come from our own test runs, which are directional and are not a share of Perplexity’s traffic. Every outside source below was opened on 7 November 2026.

Why does Perplexity cite sources at all?

Perplexity is an answer engine: it searches the web in real time, reads a handful of pages and writes one synthesized answer with numbered citations underneath. Its Help Center calls this source transparency, but it gives no rule for how sources are chosen. In our 150-query real estate study it listed sources on 142 of 150 answers, against 135 of 147 for ChatGPT and 55 of 150 for Gemini, which makes it the engine where the route to a citation is easiest to read.

Source: CiteVantage, 150-query real estate AI search study and the domains behind it, fieldwork 8 to 12 September 2026, recounted from the published CSV on 7 November 2026. Gemini returned a source list on a minority of answers, so its domain count is a floor.

Two things follow from how it works. It reads the live web, so crawlability matters more than on an engine that answers from memory. Our working assumption, which Perplexity has not confirmed, is that a claim repeated across several trusted pages is safer for it to cite than a lone assertion. The width is the point most people miss. Perplexity cited about twice as many distinct domains as ChatGPT, and only 68 of ChatGPT’s 305 cited domains also appeared among Perplexity’s 618, so a mid-market site with a sharp page has a real shot even against bigger domains. That is also why Perplexity is a productive engine for independent agents and small teams, and it is one of the engines our real estate AI visibility audits check.

What are the 9 source signals Perplexity rewards?

Perplexity does not say what it rewards, so “rewards” here means a pattern that shows up in a study we could read or in our own tests. We group the signals into four tiers by leverage: extraction, recency, authority and structure. Work them in that order, because a page Perplexity cannot extract cleanly will not be cited however authoritative the domain behind it is.

You don’t need all nine on day one. Get crawler access and an answer-first capsule live first, because those two unblock everything else.

How do you write content Perplexity will extract?

Lead with the answer. Every section should open with the direct response in one or two sentences, then expand. We found no published test of answer position on Perplexity, so this is a house rule, and Google’s guide to AI features says you don’t need to write in a specific way just for generative AI search. We follow the rule because a person skimming the page wants the answer first, and a passage-lifting engine has the same preference.

The pattern we use on every page:

• Open with a 40 to 60 word capsule that answers the page’s core question, written to be lifted whole.
• Use question-format H2s that mirror how buyers actually phrase the query.
• Keep passages self-contained so a section reads as a complete answer without the rest of the page.
• Drop in numbers, dates and named specifics Perplexity can quote, and use tables for any comparison or dataset.

Write definitively, because hedged copy (“there are many factors that may potentially influence”) gives the model nothing to lift.

Does Perplexity favor fresh content?

Somewhat, and less than ChatGPT does. In Ahrefs’ July 2025 study of 16.975 million cited URLs, the content cited in Perplexity answers averaged 1,166 days since publication against 1,416 for organic results in its table, while ChatGPT’s citations averaged 958 days. Ahrefs adds that cited content still averages about 2.9 years old, so a page does not need to be new to be cited.

That makes freshness the cheapest signal to fix and a weaker one than the table in most guides implies. Put a visible “last updated” date on the page and not just in metadata, refresh your top pages every quarter with real changes and current data, and replace a stat or screenshot when it ages out. Ahrefs also repeats Google’s warning against changing a date without changing the page, so treat your best pages as living documents and edit them for real.

Why do Perplexity citations depend so much on earned media?

Because the sources it cites are mostly other people’s sites. Our assumption is that Perplexity treats agreement across reputable third-party pages as a safer basis for an answer than a claim on your own blog, though it has not said so. Past the ceiling for owned content, the work moves into the publications, directories and review sites Perplexity already cites.

46.7%

of the citations to Perplexity’s ten most-cited domains are Reddit, in Profound’s analysis of 680 million citations.

That is why Reddit punches so far above its domain authority. Reddit makes up around 46.7% of the citations to Perplexity’s ten most-cited domains in Profound’s analysis of 680 million citations from August 2024 to June 2025, and Wikipedia is not in that top ten. That is Reddit’s share of the top-ten list, and in the same data Reddit’s share of all Perplexity citations was 6.6%, so it is the largest single source without being most of the answer. Specific comments that answer a real question in an active subreddit are the version worth posting, and real advice with numbers and outcomes beats marketing copy by a wide margin.

Where to build authority that moves citations:

• Earned editorial mentions, such as a quote in a trade publication, an industry roundup or a podcast writeup. The durable signal.
• Helpful, specific contributions in the communities your buyers read, Reddit and niche forums included.
• Listings and roundups on the domains Perplexity already cites for your category.
• Original data you publish first, whether a survey, a benchmark or a first-party study, which gives other sites a reason to cite you.

We publish studies with the row-level data attached for that reason, such as the 500-query Houston real estate study. We have not run a controlled test of how much original data moves Perplexity specifically.

Is Perplexity SEO the same as Google SEO?

No, though the two overlap more than ChatGPT and Google do. In Ahrefs’ August 2025 study of 15,000 long-tail prompts, 28.6% of Perplexity’s cited URLs ranked in Google’s top 10 for the same prompt, against 8.0% of ChatGPT’s in-text citations. Ahrefs notes that cited pages need not rank for the original query, because assistants search variations of it. Ranking still helps Perplexity find you, and extraction and corroboration decide what happens after that.

Perplexity writes one answer and footnotes several sources, so being “page one” matters less than being extractable and trusted. The engines also barely share sources: only 68 of ChatGPT’s 305 cited domains appeared among Perplexity’s 618 in our study, and 83% of all cited domains reached one engine of five. A page that wins on ChatGPT can be invisible on Perplexity, so you measure each engine separately.

For the other engines, see our guides to getting cited by ChatGPT, how to rank in ChatGPT, getting cited by Google Gemini and optimizing for Google AI Overviews. The underlying discipline is generative engine optimization, and the split between GEO and classic search is in GEO vs SEO vs AEO.

What is the difference between PerplexityBot and Perplexity-User?

PerplexityBot builds the index Perplexity searches and links from, and Perplexity says it is not used to crawl content for AI foundation models. Perplexity-User fetches a page when a user’s question needs it and, per Perplexity, generally ignores robots.txt because a user requested the fetch. The control point for Perplexity-User is your firewall, not your robots.txt.

Perplexity’s crawler documentation (source of the table) says to pair each bot’s user agent with that bot’s own IP ranges in a firewall rule, and to put the allow rule above any rule that blocks or challenges traffic. It also warns that Cloudflare’s Bot Fight Mode cannot be skipped with WAF custom rules and can still challenge matching bots.

A block can also hit one crawler and miss the rest. On our own host, PerplexityBot passed on 8 of 8 requests on 14 September 2026 while the full GPTBot string was refused on 8 of 8, as the GPTBot case study shows. You can run the same kind of test on your own site by requesting a page with each user agent above and a cache-busting query string, for example curl -I -A "<user agent>" "https://yoursite.com/?cb=1", and changing the number each time. That shows what a rule keyed on the user agent does. It cannot show how a rule keyed on Perplexity’s published IP ranges treats the real bot, which is why the IP lists matter.

Robots.txt states a preference and the firewall enforces it. Cloudflare reported on 4 August 2025 that, on new domains set to disallow all bots, Perplexity still returned detailed information about their content, and that it saw a crawler with a generic Chrome-on-macOS user agent from addresses outside Perplexity’s published ranges. Cloudflare’s post quotes no reply from Perplexity, so treat it as Cloudflare’s account.

What about llms.txt, JSON-LD and other myths?

Skip llms.txt for citations. Google says its Search ignores such files, Perplexity’s crawler documentation mentions llms.txt only as a pointer to its own docs, and in a 12-week log panel of 83 sites that ran to 19 July 2026 PerplexityBot fetched robots.txt 775 times and llms.txt never. We cover what llms.txt does and which engines read it separately.

JSON-LD is overrated as a citation lever too. Schema helps engines parse your page, but a SearchVIU experiment reported by Evil Martians found ChatGPT, Claude, Perplexity, Gemini and Google AI Mode all missed product data that existed only inside JSON-LD, and the write-up gives no date or method. Keep your schema, and write the same facts in plain, visible body text where the model reads.

Quick myth check:

• llms.txt: no measured citation benefit today. Don’t prioritize it.
• JSON-LD alone: helps parsing, won’t carry a fact the body text omits.
• Blocking AI crawlers “to protect content”: keeps PerplexityBot from surfacing your pages in Perplexity search.

How does Perplexity cite real estate agents and brokerages?

Perplexity names more real estate businesses than the other engines we measured, and it leans on portals and referral networks to find them. In our 500-query Houston study it named 1,942 distinct firms on the 432 questions all four surfaces answered, against 803 for ChatGPT, and sent 29.6% of its citations to property portals and 10.4% to agent referral networks.

Houston study

1,942

distinct firms named by Perplexity on the 432 questions all four surfaces answered, against 803 for ChatGPT.

Photo: Nicole Seidl, Pexels

Width cuts both ways for an agent. A long list means more open places on each answer, but many of Perplexity’s picks never carry over to the other engines. In our 150-query study, 55% of Perplexity’s citations went to domains no other engine cited, the highest exclusive share of the five, as the open list of domains AI cites on real estate questions shows. Being named by Perplexity in one suburb also says little about the next suburb over, since 82.6% of the Houston agents who were named appeared in just one of 40 areas.

The practical version for a team or brokerage is to run the area question in Perplexity and read the footnotes. Where a portal page or referral network is cited, a complete profile there is the quickest fix within your control. Where a local guide or news site is cited, an answer-first area page of your own, with current figures, is the page to build. Our GEO for real estate programme works through that list area by area, and the wider playbook for agents is in AI search for real estate agents.

How to get cited by Perplexity: the audit to run today

Five steps, no tools and about thirty minutes. The point is to find out whether Perplexity cites you on the questions your buyers ask and, if not, which sources it used instead. Use a private window and stay logged out, because an account carries your history into the answer. Write the answers down, because you will re-run this later:

1. Ask Perplexity your buyer’s real question, for example “best real estate agent in [area] for first-time buyers.”
2. Read the numbered sources under the answer. Write down every domain.
3. Check whether you’re there. If not, note who is and which specific page got cited.
4. Match that page’s format, answer-first and freshly structured, then earn a mention on two or three of the cited domains.
5. Re-run the same query in three to four weeks and compare.

Answers change between runs, so ask each question on three different days before you read anything into one result. Do that across your ten most important buyer questions and you have a citation gap map. That map is exactly what we build in a free AI visibility audit across ChatGPT, Perplexity, Gemini and Copilot, scored 0 to 100, so you can see who Perplexity cites instead of you.

Start with the two unblockers: allow PerplexityBot, and put an answer-first capsule at the top of your most important page. Then layer in earned media and keep the page maintained. How to rank in ChatGPT covers the same checklist for ChatGPT.