How to Get Cited by Perplexity

How Perplexity actually picks its sources

Perplexity is not a search engine that shows ten links. It is an answer engine. When someone types "who does emergency AC repair in Fort Myers," it fires a real-time web search, pulls a handful of pages, extracts the passages that answer the question, and writes a paragraph with little numbered chips next to it. Each chip is a citation to a page it used. Your goal is to be one of those pages.

Three things decide whether your site makes the cut. First, retrieval: your page has to surface in the search Perplexity runs behind the scenes, which leans on Google-style and its own index signals. Second, relevance: the passage on your page has to directly answer the question, in plain language, near the top. Third, trust: Perplexity favors sources that other places corroborate, so a shop with a consistent name, address, and phone across the web reads as more real than one that appears out of nowhere.

The part contractors miss is that Perplexity reads for extractability. It wants a sentence it can lift and attribute. A homepage that opens with a hero slogan and a stock photo gives it nothing to quote. A service page that opens with "We provide 24-hour emergency air conditioning repair in Lee County, typically on-site within two hours," hands it a citable answer on a plate. Same business, very different odds of getting named.

Perplexity also cites more sources per answer than most people assume, often four to eight links for a local-services question. That is good news. You are not fighting for one slot. You are trying to be clear and trustworthy enough to make a short list.

One more thing to understand: the answer engine reads recency and specificity as trust signals. A page that names a real service area, quotes a real range, and reads as maintained this year beats a vague page that could have been written for any city in any decade. When two shops are otherwise equal, the one that commits to concrete detail is the one the model feels safe attributing an answer to. Contractors tend to hedge on their own sites out of habit. On the citation layer, hedging costs you the slot.

Build source pages that answer one question cleanly

The unit of AI citation is the answer passage, not the page. Perplexity lifts a specific sentence or short block and attributes it. So structure every page that matters around one clear question and answer it in the first hundred words, before any marketing.

Lead with the direct answer. If the page targets "how much does a roof replacement cost in Tampa," the first paragraph should state a real range and the factors that move it, in complete sentences. Then expand. This inverts how most contractor pages are written, where the answer, if it exists, is buried under three paragraphs of "why choose us."

Use structure the model can chunk. Question-style H2 and H3 headings, short paragraphs, and a tight list or table for anything with discrete parts. A pricing question wants a table. A "what are the signs I need a new water heater" question wants a numbered list. These formats are easy for a language model to extract and cite verbatim.

• One page, one question. Do not bury the answer to "emergency service response time" inside a general services page. Give it room.
• Complete sentences. Fragments and bullet-only pages are hard to quote. Write at least one clean, standalone sentence that states the answer.
• Real specifics. Named service areas, real ranges, actual timeframes. Vague copy gets skipped for a source that commits to a number.
• Current and dated. Perplexity favors fresh pages. A page that reads as maintained beats one that looks abandoned.

Watch out for the two habits that quietly kill extractability. The first is the hero-slogan opener, where the top of the page is a headline and a photo and the actual answer is three scrolls down. The model reads top-down and weights the opening heavily, so an answer buried below the fold might as well not exist. The second is answering only in fragments and pull-quotes. A snappy "24/7. On call. Always." strip looks nice to a person and reads as noise to a machine that needs a complete, attributable sentence. Give the model at least one full sentence it can lift clean.

This is the AEO side of the work: answer-engine optimization. You are writing pages a machine can read out loud and attribute to you. That is a different craft than stuffing a keyword and hoping a blue link ranks. The blog engine that produces those pages at volume is its own service; here we care about the shape of the answer, not the publishing pipeline.

Mark up your pages so a machine can parse them

Structured data is how you tell a machine, in its own language, what your page is and what facts it contains. Perplexity and the search layer it rides on read schema to disambiguate a business and to pull clean facts without guessing. For a contractor, three types carry most of the weight.

The FAQPage type earns its keep for citations specifically. Each entry is a clean question paired with a clean answer, structured so an answer engine can extract it whole. If a homeowner asks Perplexity a question you have answered in your FAQ schema, you have handed it exactly what it wants to quote.

Get the details right. Your NAP (name, address, phone) in the schema has to match your visible page and your listings elsewhere, byte for byte. A phone number that differs by a formatting quirk between your schema and your Google profile plants doubt about whether they are the same business. Validate every page with a structured-data testing tool before it ships. Broken JSON-LD is worse than none, because it can get your markup ignored entirely.

A word on how much schema to write. You do not need to mark up every element on the page. You need the facts that answer questions marked up cleanly: who you are, what you do, where you do it, and the specific question-answer pairs a homeowner cares about. Over-stuffing schema with claims your visible page does not support is a fast way to get your markup distrusted or ignored. Match the schema to the page. Say only what the page says.

Schema for AI citation is what we own here. Schema that also drives Google rich results or map-pack ranking overlaps, but the version tuned so an answer engine can parse and quote you is the canonical job of this work.

Earn the third-party corroboration Perplexity trusts

Perplexity does not take your word for it. When your site claims you are a licensed electrician in Sarasota, the model is more confident naming you if independent sources agree. That corroboration is a big part of why one contractor gets cited and an equally good one does not.

The sources that move the needle for local trades are the ones an answer engine already reads and trusts. Directories and profiles with your consistent NAP. Your Google Business Profile. Industry and licensing databases. Chamber and association listings. Reviews on platforms the model indexes. None of these are exotic. What matters is consistency: the same business name, the same address format, the same phone, everywhere. Conflicting information is the fastest way to get skipped, because the model cannot tell which version of you is real.

1. Lock your NAP everywhere. One canonical spelling and format. Fix the old listings that still show a former address or a disconnected number.
2. Claim and complete the obvious profiles. Google Business Profile, the big directories, your trade's licensing lookup. Fill every field.
3. Get named on pages you do not own. A supplier's dealer locator, a local news mention, an association member page. Third-party pages that name you are strong corroboration.
4. Keep reviews flowing. Fresh, real reviews signal an active, real business, which the model reads as trustworthy.

Think of it the way an inspector thinks about a permit. Your site is the claim. The corroboration is the paperwork that says the claim checks out. An inspector who can pull a matching permit signs off without a second thought. An answer engine that can pull matching listings names you without hedging. When the paperwork conflicts, both of them slow down and look elsewhere.

The map-pack and "near me" proximity side of Google Business Profile is a ranking discipline of its own and lives in local SEO. Here we care about the profile purely as a corroborating source: does it exist, is it consistent, does it confirm what your site claims. That confirmation is what lets Perplexity name you without hedging.

Track whether Perplexity is citing you

You cannot manage what you do not measure, and AI citations do not show up in your normal analytics the way a Google click does. So you have to check on purpose. The method is not glamorous. You ask the questions your customers ask and you read the answers.

Build a short list of the real queries a homeowner in your area would type: "best roofer in [your city]," "emergency plumber near [your area] open now," "how much does [your service] cost in [your region]." Run each one through Perplexity and note three things. Does your business get named in the written answer? Does your site appear as a numbered citation? And which competitors are getting cited instead of you?

• Named vs. cited. Being cited (your link in the sources) is the win. Being named in the prose is even better. Track both.
• The competitor set. Whoever Perplexity cites for your money questions is who you have to out-clarify. Read their pages. See what they answered that you did not.
• Referral traffic. Perplexity sends real clicks. Watch your analytics for referrals from its domain, and check whether AI-referred visitors behave differently than search visitors.
• Recheck on a cadence. Answers shift as pages get re-crawled. A monthly pass catches whether your fixes are landing.

Keep a simple log. A spreadsheet with your query list down the left and a monthly column for each check is enough. Mark whether you were named, whether you were cited, and who else showed up. Over a few months that log tells you a story your gut cannot: which fixes moved which queries, which competitors are working the channel, and where you are still invisible. It also keeps the conversation honest when you evaluate whether the work is paying off.

Expect this to take time. When you ship better source pages, clean schema, and consistent listings, the model has to re-crawl and re-evaluate before it changes its answer. A realistic window for competitive local terms is several months, not several days. The upside is durability: once you are the clearest, most corroborated source for a question, you tend to stay cited until a competitor does the same work.

What to fix first (a contractor's order of operations)

You do not have to do everything at once. There is an order that gets the most citation lift for the least motion, and it starts with the pages that already exist.

1. Rewrite your top service pages to answer, not sell. Put the direct answer to the page's core question in the first hundred words. This alone changes how extractable your site is.
2. Add clean schema to those pages. LocalBusiness on the homepage, Service on each service page, FAQPage anywhere you have real questions. Validate before shipping.
3. Reconcile your listings. One NAP, everywhere, matching your schema. Fix the stale profiles.
4. Build the missing answer pages. The specific cost, timing, and "do I need" questions your customers actually ask, one page each.
5. Measure, then repeat. Run your query list through Perplexity monthly and let the results tell you where to point next.

A note on what not to chase. There is no submission form to get into Perplexity, no paid inclusion, no trick tag that forces a citation. Anyone selling you a shortcut is selling you nothing. The channel rewards the same things a good tradesman already respects: do the job right, be exactly who you say you are, and make it easy to verify. The contractors getting cited today are not the loudest. They are the clearest.

This is the fastest-moving slice of contractor search, and most agencies still have no answer for it because they never learned the citation layer. If you already rank on Google but keep hearing "I asked ChatGPT and it never mentioned you," this is the gap, and it is fixable.