How to Get Your Business Cited by Perplexity (UK Guide)

Of the AI search tools, Perplexity is the one I recommend business owners start with, for a practical reason: it shows its working. Every answer comes with the sources it used, so you can see exactly which pages shaped the response and which businesses were named.

That transparency makes it an unusually good learning tool and, for smaller UK businesses, a realistic place to earn visibility. Here is what Perplexity says about how it works, what the research suggests, and what I would do for a London business that wants to be cited.

How does Perplexity find and cite sources?

Perplexity runs its own search crawler, PerplexityBot, which it says is designed to surface and link websites in Perplexity's search results. It states that this bot is not used to crawl content for AI foundation models, so allowing it is about appearing in answers, not about training.

A second agent, Perplexity-User, visits pages when a person asks a question and generally ignores robots.txt rules, because a user requested the fetch. Both publish IP ranges, and Perplexity gives firewall setup instructions for Cloudflare and AWS, which tells you how often sites block it by accident. I cover the settings in the post on AI crawlers.

How often does Perplexity name local businesses?

In SOCi's 2026 study, Perplexity recommended 7.4% of business locations analysed, ahead of ChatGPT at 1.2% and behind Gemini at 11%. The figures cover nearly 350,000 locations across 2,751 multi-location brands and were reported by Search Engine Land. Perplexity-recommended locations averaged 4.1 stars.

SOCi also reported that business profile information was only about 68% accurate on ChatGPT and Perplexity. That is a polite way of saying these systems frequently get basic facts wrong, and a business whose facts are consistent and well corroborated is easier for them to describe correctly.

Is Perplexity realistic for a small business?

The research suggests it is, because lower-ranked sites appear to gain more from optimisation than top-ranked ones. The GEO paper, accepted to the KDD 2024 conference, found that adding citations to a page led to a 115.1% increase in visibility for websites ranked fifth in search results, while the visibility of the top-ranked website decreased by 30.3% on average.

That is a result from a research benchmark and one set of engines tested at the time, not a promise about Perplexity today, and I cover the caveats in the post on what the GEO research found. But the direction matters for a London firm competing against bigger names: an answer engine that draws from many sources gives a well-evidenced smaller site a chance that classic search rankings often do not.

What kind of page gets cited?

A page with a specific, quotable answer, backed by evidence and a named source. Perplexity has to justify each statement with a link, so it prefers pages where the claim and its support sit together. A page that says 'we are the leading firm in London' gives it nothing to cite. A page that says 'a residential conveyancing in London typically takes eight to twelve weeks, and here is why' does.

• Specific numbers with a source or a basis, such as a time range, a price range or a count from your own work.
• Definitions and direct answers in the first sentence under a clear heading.
• Original information: your own data, a worked example, a comparison you ran yourself.
• A named author with credentials, and a visible last-updated date.
• Honest limits and caveats, which make a claim more credible, not less.

Notice that none of this is Perplexity-specific. It is the same non-commodity standard Google describes, and I go through it in the post on commodity versus non-commodity content. A page good enough to be cited by one answer engine tends to be good enough for the rest.

Which outside sources does Perplexity draw on?

Vendors report that answer engines lean on discussion sites, video, reference sites and directories, but I would treat any specific list as unverified. One AI visibility vendor's comparison names Reddit, Quora, YouTube, Wikipedia and industry directories as platforms these tools trust. That is a vendor claim, and it will differ by query.

The way to replace that guess with evidence is to look. Run your prompts in Perplexity and read the sources panel, as in how to check your AI visibility. The sources it cites for your category are the sources that matter for your category, and they are often a short list of directories, trade publications and review sites you can act on.

• Directories and registers in your sector, kept accurate and complete.
• Trade body and regulator pages where you are listed.
• Local and trade press you can legitimately earn coverage in.
• Review platforms your customers actually use.
• Community discussions where you can give genuinely useful, honest answers without promoting yourself.

How do I use Perplexity to research my competitors?

Run your prompts and list the domains it cites, then work out which of those you could appear on. Because the sources are shown, the citation list is a map of the pages shaping answers in your category. A pattern usually emerges within a dozen prompts: a few directories, a couple of trade publications and one or two comparison sites account for most of it.

Keep a note of what you learn about the sources, because it is reusable. The same directories and publications that Perplexity cites for your category often feed other assistants and Google as well, so time spent getting an accurate, complete entry on a handful of them pays back across several surfaces. It is also the kind of work a human customer checking you out will notice, which is a better reason to do it than any algorithm.

What does success look like?

An appearance rate that rises across your prompts, accurate descriptions, citations to your own pages, and enquiries you can trace. Anything less specific is a feeling. Set the baseline before you change anything, repeat the same prompts monthly, and record the three numbers that matter: how often you are named, how often your own pages are cited, and how often the description is right.

Give it a quarter before you judge. Answers vary between runs, citations take time to build and the sources Perplexity draws on do not update overnight. A change you can see across several prompts and several runs is real. A single good answer is a coincidence.

What should I avoid?

Do not manufacture mentions, and do not write for the bot at the expense of the reader. Google's own guidance says seeking inauthentic mentions is not as helpful as it might seem, and while that is Google speaking about Google, the incentive to discount manufactured signals applies to any answer engine that wants to be trusted.

• Bought or fake reviews. They are also unlawful for UK businesses, which I note in the post on reviews.
• Astroturfed forum posts. If a community spots promotion it will remove it, and the footprint can harm you.
• Keyword stuffing. The GEO research found it offered little to no improvement on generative engine responses.
• Blocking the crawler by accident through a firewall rule you forgot about.

What would I do in the first fortnight?

1. Open Perplexity and run ten prompts your customers would ask, noting who is named and what is cited.
2. Confirm PerplexityBot is allowed in robots.txt and your firewall.
3. Read the cited pages for your category and note what they have that you do not.
4. Add one specific, sourced answer to each of your three key pages.
5. Make your business facts identical across the site, profile and the directories Perplexity cites.
6. Plan one piece of original data you can publish, even a small one.

Then repeat the prompts a month later. Because answers vary between runs, look for a change in the pattern rather than a single result. If you would like the audit done for you, my AI search visibility service begins exactly there.