Getting recommended by ChatGPT and AI Overviews, included

Why do we need to say this plainly, up front?

One UK business owner posted that their SEO agency, already being paid four figures a month, wanted a separate fee on top for ‘AI visibility’. The response from other business owners and SEO practitioners was blunt: it was called a scam, and the advice was to keep the money where it already sat.

They were right to be sceptical. AI search tools like ChatGPT and Perplexity largely rely on the same underlying web index that Google uses. There is no separate hidden system to unlock with a bigger invoice. Getting recommended by an AI answer depends on almost exactly the same fundamentals as ranking well in ordinary search: a page that answers the question clearly, business details that are consistent everywhere, and other real sites and reviews backing you up.

That overlap is why AI visibility is included in both FlyRank retainers at no additional cost, and always will be. If we ever thought it genuinely required separate, additional work, we’d tell you honestly rather than quietly adding a line to your invoice.

What actually moves the needle in AI answers?

Independent research into what gets cited by AI answers points at the same handful of things, and none of them are exotic. Pages that already rank well in ordinary Google search are heavily overrepresented among AI citations — being findable the normal way is close to a prerequisite. Page length barely matters; over half of pages cited by AI tools are under a thousand words, so padding a page doesn’t help. Freshness does matter — AI tools favour pages that have been updated recently over pages that were republished as a duplicate. And mentions of your business elsewhere on the web — reviews, local press, directories, forum discussions — correlate strongly with how often AI tools recommend a brand.

In practice that means: keep your Google Business Profile accurate, keep your website’s service pages clear and current, get genuine reviews, and get mentioned by other real, relevant sites. That is also, almost word for word, what good local SEO already does.

• Clear, current service pages that answer real customer questions directly
• Consistent business name, address and phone number everywhere online
• Genuine reviews and other sites mentioning your business
• Pages kept up to date rather than republished as near-duplicates
• Structured data that helps machines, not just humans, read your pages

What won’t we sell you?

We won’t sell you an ‘AI Visibility Score’ with an invented number behind it, and we won’t promise a specific percentage lift, because no honest agency currently has UK data solid enough to back that up. The one paid product we do offer that’s specifically about AI search is the AI Visibility Sprint, a one-off, defined piece of work for a business that wants a focused push rather than an ongoing retainer — priced and scoped upfront, not billed as a silent monthly addition.

How do we check where you stand today?

We ask ChatGPT, Google’s AI Overviews and Perplexity the questions your customers are likely to ask, and note who gets recommended and why. That’s the same free visibility check available from the homepage, and it’s the starting point for any AI search work we do.

How are AI answers actually built?

It helps to understand the mechanism, because it explains why there’s no separate trick to sell. ChatGPT, Perplexity and Google’s AI Overviews mostly draw on the same underlying web index that ordinary search already uses — they aren’t running a second, secret crawl of the internet. Ranking well in normal Google search is close to a prerequisite for being cited in an AI answer, not a separate achievement.

That’s also why the two names you’ll sometimes see — answer engine optimisation and generative engine optimisation — describe the same underlying work as SEO with a newer label attached, rather than a distinct discipline requiring a separate budget.

Why do comparisons and lists get cited more?

Pages formatted as a list, a comparison or a genuine review are disproportionately cited by AI tools, compared with plain descriptive pages. That’s part of why the FlyRank blog leads with comparisons and roundups rather than only single-topic articles — it’s a format AI tools are already inclined to pull from, not a trick.

Why do we update pages instead of republishing them?

Recently updated pages are cited more often than stale ones, even when the underlying content barely changes. So we update pages in place — correcting, adding, refreshing — rather than publishing a near-identical new version every few months to chase a freshness signal. It’s a more honest approach and, as it happens, the one that performs better.

What’s the honest, deliverable version of this work?

There is a genuine, deliverable version of AI search work beyond what ordinary SEO already covers: making sure the citations that sit outside Google’s own ecosystem — trade bodies, niche directories, industry round-ups — are accurate and present, because those seem to matter disproportionately to how ChatGPT and Perplexity treat a business. It also means making sure your business gets talked about naturally across the web: genuine reviews, community mentions, forum answers, not manufactured content designed to look organic.

That’s real, specific work, and it’s exactly what’s already included in Runway and Lift under our method — it isn’t a reason to add a second invoice.

What are we honestly unsure about?

We won’t tell you a precise percentage of your customers are already discovering you through AI search, because no one has solid, independent UK data on that yet for a business your size — anyone quoting you a confident number for that is likely working from a vendor’s marketing page, not measured fact. What we can show you is what specifically comes back today when we ask the AI tools the questions your customers might ask, which is a real, checkable answer rather than a guess dressed up as a statistic.

AI search vs SEO — is there really a difference?

In practice, no, not for a local business. The overlap is close enough that a UK SEO practitioner replying to the surcharge thread mentioned earlier on this page put it plainly: SEO and answer engine optimisation aren’t too dissimilar, and both require good site structure, customer-focused pages and trustworthy links. Where a genuine difference shows up is in emphasis rather than technique — AI tools seem to weight off-site mentions and citations slightly more heavily than a plain ranking algorithm does — but the day-to-day work is the same list of jobs under the SEO page, not a second list.

Month by month, how does AI visibility move?

Month one is largely diagnostic: we check what ChatGPT, AI Overviews and Perplexity currently say about your business and your competitors, and note the gaps — missing citations, thin or inconsistent business information, an absence of genuine mentions elsewhere. Runway work (already covered in the SEO retainer) fixes the consistency issues that affect both ordinary rankings and AI answers.

By months two and three, the content and outreach work from Lift starts to feed both channels at once — a genuinely useful service page or a real mention on a trade directory helps your Google ranking and your chances of being cited in an AI answer for the same reason. We don’t run a separate AI-specific campaign alongside it; the same monthly report covers both.

From there it’s ongoing: rechecking what the AI tools say periodically, and folding anything that’s changed into the next month’s plan. If a competitor starts appearing in AI answers where you don’t, we’ll flag it and explain why.

What do we deliberately not do here either?

We won’t sell an ‘AI Visibility Score’ as a headline metric, because no scoring system in this category is independently verified, and presenting an invented number as data would be exactly the kind of thing this page argues against. We won’t guarantee a specific AI tool will recommend you, for the same reason we won’t guarantee a Google ranking. And we won’t run this as a project separate from your SEO work, because doing so would imply it’s a different discipline requiring a different invoice, which it isn’t.

Which businesses gain the most from this right now?

Businesses where the buying decision involves research before a call — clinics, professional services, higher-value trade work like an extension or a full rewire — tend to see AI tools used more often in that research phase than a straightforward, immediate-need search like ‘plumber near me now’. That doesn’t mean it’s irrelevant to urgent-need trades; it means the content that helps there is slightly different — clear, specific service pages that answer exactly what’s involved, rather than long explanatory guides nobody in an emergency has time to read.

What would make us change this approach?

If AI tools genuinely started requiring separate, additional work beyond good SEO — a distinct technical format, a paid submission process, anything that couldn’t be folded into the existing retainer — we’d tell you plainly and explain what changed and why, rather than quietly adding a line to an invoice. Nothing in the evidence we can see today points that way, but we’re not going to pretend the landscape is frozen either.