Perplexity SEO: How Perplexity Finds and Cites Business Websites
Perplexity SEO is the work of making sure Perplexity can reach your website and has something on it worth citing. Perplexity searches the web while it writes each answer and lists numbered sources beside the text, so your site either shows up in that list or it doesn't. You control two things: whether its crawler can get in, and whether your pages state the facts a question needs. You don't control which sources it picks.
How Perplexity finds and cites sources
Perplexity answers by searching the web at the moment you ask, then writing a reply with citations. Its help center says content is sourced from the web in real time and that each response includes citations and links to the original sources.
That's different from a chatbot answering from memory. The question "who installs paver patios in Fort Collins" goes out as a search, Perplexity reads what comes back, and the answer cites the pages it used. If your site isn't in what it can read, the answer falls back on directories, review sites, and competitors.
PerplexityBot vs Perplexity-User
Perplexity runs two separate agents, and they follow different rules. Its official bot documentation describes them like this:
- PerplexityBot crawls the web to surface and link sites in Perplexity's search results. It is not used to gather content for training AI foundation models. It follows robots.txt.
- Perplexity-User visits a page when a person's question calls for it, and may link that page in the answer. Because a user triggered the visit, Perplexity says this fetcher generally ignores robots.txt rules.
The practical upshot: robots.txt decides whether your pages get into Perplexity's search results through PerplexityBot. It does not reliably stop a live fetch by Perplexity-User. Perplexity also notes that robots.txt changes can take up to 24 hours to apply.
What Perplexity SEO means for a local business
For a design studio or a landscaping company, Perplexity SEO comes down to three jobs: open the door, put the facts in text, and keep other sites consistent with yours.
Here's how the control splits.
What you control:
- Whether PerplexityBot can crawl your site, through robots.txt and your firewall.
- What your pages say, how specific they are, and whether each service has its own page.
- Your listings and profiles elsewhere, which Perplexity may cite instead of you.
What you don't control:
- Which sources Perplexity chooses for a given question.
- How it words the answer, or whether it names you in the text.
- Live fetches by Perplexity-User, which don't depend on robots.txt.
Perplexity's bot documentation explains access. It says nothing about a way to rank higher inside answers, and we'd be skeptical of anyone selling a "Perplexity ranking" formula.
Check whether Perplexity can reach your site
Start here, because a blocked crawler makes everything else pointless. Small business sites get blocked by accident more often than on purpose, usually by a hosting setting nobody remembers turning on.
- Open yourdomain.com/robots.txt. Look for a rule naming PerplexityBot, and for a blanket "Disallow: /" under "User-agent: *" that would catch it too.
- Check your hosting dashboard or CDN for "bot protection," "block AI crawlers," or similar switches. Some are on by default.
- If you use a web application firewall, Perplexity advises allowing its bots by matching both the user agent and the source IP, using its published lists for PerplexityBot and Perplexity-User. Those lists change, so a rule written once and forgotten can go stale.
- If you can see server logs, search for "PerplexityBot" in the user agent field. Visits mean the crawler is getting through. A string of blocked responses means it isn't.
- After any fix, give it a day before testing. Perplexity says robots.txt changes can take up to 24 hours.
If you'd rather keep Perplexity out entirely, a robots.txt rule for PerplexityBot covers its search crawler. Just know you'll also drop out of the source lists it builds from that crawl.
Pages Perplexity can actually use
Perplexity cites pages, not websites. A question about a specific job needs a specific page with the answer in plain text.
Think about the shape of the questions people ask it. They tend to be longer and more specific than a Google search: "landscape designer near Boulder who works with native plants and can handle a sloped backyard." An answer to that needs a page that mentions native plants, slopes, and Boulder. A homepage that says "beautiful outdoor spaces" doesn't qualify.
A few page types that give Perplexity something to cite:
- A service page for each kind of job, with the towns you cover written in the text.
- Project pages that name the place, the scope, and the problem solved, beyond the photos.
- A plain pricing or "what it costs" section with your real starting point and what moves it.
- An About page that states who runs the business, since questions often ask "who is" before "who does."
Say you run a small interior design studio in Minneapolis that does a lot of basement conversions. If that phrase appears nowhere on your site, Perplexity has no page of yours to cite for it, however good your photos are. One project page titled and written around a basement conversion changes that.
For the wider picture of how other assistants handle sources, see our guide to checking how AI assistants describe your business, which compares ChatGPT, Claude, and Google side by side.
How to test your business in Perplexity
Perplexity is the easiest AI tool to audit, because it shows its sources. The test below is built around that source list as much as the answer text.
Run the questions
Pick five questions a new client might ask, with your town in them. Include one that names your business directly, such as "What does [studio name] in Minneapolis specialize in?" Run each in a fresh session, then ask one natural follow-up in the same thread, since many people do.
Read the source list, not only the answer
For every answer, open each numbered citation and sort it into one of four groups:
- Your own site, and which page.
- A directory or review site with a profile of you.
- A competitor's site.
- A general article that names nobody local.
This tells you more than whether your name appeared. If Perplexity cites a directory profile of you instead of your site, that profile is now speaking for you, so it had better be correct. If it cites your homepage for a service question, you probably need a dedicated service page. If only competitors and generic articles show up, the gap is on your side.
Check the facts it repeats
Compare every fact in the answer to the truth: address, service area, services, price hints, years in business. Trace each wrong one to its cited source. Fix your own page first, then the outside profile. If an answer is plainly wrong, Perplexity's help center says you can report it with the flag icon below the answer.
Repeat on a schedule
Answers vary between runs, so keep the same five questions and rerun them monthly. Write down the date, the cited pages, and any errors. A short log beats a memory of "I think we showed up last time."
Should you block Perplexity?
That depends on what you want from it. Blocking PerplexityBot keeps your pages out of the results it builds from its crawl, but Perplexity says it doesn't use that bot to train foundation models, so blocking it doesn't protect your content from training. For a local service business that wants inquiries, we'd keep PerplexityBot allowed.
The same logic applies to other assistants, and each documents its own crawlers. ChatGPT is covered in our guide on getting cited by ChatGPT.
Where your website fits in
Most of what moves Perplexity results for a small studio is on the site itself: separate service pages, project pages with real text, and facts that match your profiles. That's the kind of site we build for interior design studios and landscaping companies, with copy written from your answers to our questions and checked by you before launch. See what's included in our AI SEO services. Projects start from $1,000.
FAQ
If Perplexity keeps citing everyone but you, the fix often starts with the pages your site doesn't have yet. Discuss your project.
Discuss your project