How to Get Your Small Business Mentioned by ChatGPT (Not Just Ranked by Google)

The short version: Getting your small business mentioned by ChatGPT, Perplexity or Google's AI Overviews has almost nothing to do with traditional SEO and everything to do with being specific, quotable, and structured like an answer rather than a sales page. Most small business websites will never get cited because there is nothing on them worth quoting. Fix that, and you can show up in answers your competitors never see.

What made me pay attention to this

I check my referral traffic every Monday morning, out of habit more than anything. For years that report was boring: Google, a bit of LinkedIn, some direct traffic from people typing my name in. Then in January 2026 I noticed something in the "other" category I'd normally ignore. Perplexity and ChatGPT were sending me visitors. Not hundreds, but real people landing on specific pages, on specific paragraphs, clearly having been given my name as part of an answer to a question they asked an AI tool.

In the first three weeks of that month it was 41 visits from those sources combined. The same three weeks the year before, it was 3. That's not a huge number in absolute terms, but the direction matters, and the pages that got picked weren't my homepage or my services page. They were the specific, numbers-heavy posts, the ones where I'd written about what happened when I published 569 blog posts in 45 days and Google only indexed a small fraction of them. AI tools like specifics. They pulled the exact number, 5 percent, into an answer about AI content and indexing, and my name came along with it.

That taught me something that most content about "AI search optimisation" glosses over: this is not keyword optimisation with extra steps. It's a completely different game, and most small businesses are playing the wrong one.

Ranking and being recommended are not the same game

Google ranking rewards relevance, authority signals, backlinks, and increasingly, whether people click and stay. You can win at that game with decent on-page SEO, consistent publishing, and a bit of digital PR.

Being recommended by an AI tool is different. A large language model (you can read the technical definition on Wikipedia if you want the full explanation) doesn't rank ten blue links and let a human pick. It reads through what's available, decides what best answers the specific question it was asked, and writes a synthesised answer, sometimes naming a source, sometimes not. It's choosing what's quotable, not what's optimised.

That means a page stuffed with the phrase "best AI consultant London" twenty times does nothing for you here. A page that says "in 2025 I charged a client 3,200 pounds for a six week AI implementation project and here's exactly what we did in week one" gets picked up, because it's specific, it's a real number, and it answers a real question someone might ask an AI tool: "how much does an AI consultant cost."

The uncomfortable bit nobody wants to say out loud

Here's the part most posts on this topic avoid because it's not a nice thing to tell small business owners who've just spent money on a shiny new website: most small business sites have nothing on them worth an AI tool quoting. Not because the business is bad, but because the website is generic. It says "we provide bespoke solutions tailored to your needs." It has an about page that says the founder is "passionate about excellence." None of that is quotable because none of it is a fact.

I've reviewed a lot of small business sites over the years, and the pattern is consistent. The services page describes what the business does in the same vague language every competitor uses. The case studies, if they exist at all, don't name real numbers. The FAQ, if it exists, answers questions nobody asked. An AI tool scanning that page has nothing to lift. It's not that the algorithm dislikes you. It's that there's simply nothing there to cite, so it moves on to the competitor down the road who happened to publish one paragraph with a real figure in it.

This is uncomfortable because it means the fix isn't a technical plugin or a meta tag. It's going back through your own site and being brave enough to write down actual numbers, actual client names (with permission), actual mistakes and what they cost you. That's harder than installing something. It's also the only thing that works.

The seven-step audit I run on my own site

This is the process I use, roughly once a quarter, on my own website and when I'm working with clients through AI consultant work for small businesses.

• Step 1: Ask the AI tools yourself. Open ChatGP, Perplexity, and Google's AI Overview and type the exact questions your customers would ask. "Best AI tools for a small accountancy firm." "How much should a freelance marketer charge in 2026." Note who gets mentioned and who doesn't.
• Step 2: Search your own name and business name in each tool. If nothing comes back, that's your baseline. Zero mentions is normal for most small businesses starting this. It's not a failure, it's a starting point.
• Step 3: List every page on your site that contains a real number. A price, a percentage, a timeframe, a client result. If you can count these pages on one hand, that's your problem, not your marketing budget.
• Step 4: Rewrite your three most important pages so every claim has a number attached. Not "we've helped many businesses grow" but "we've worked with 34 small businesses since 2023, average project length nine weeks."
• Step 5: Add a genuine FAQ section to your service pages, written in the actual words customers use, not marketing words. Answer each one in two or three sentences that could stand alone as a quote.
• Step 6: Publish at least one piece of first-hand experience content a month. Not generic advice, an actual thing that happened, with a number in it. This is the same principle I wrote about when I looked at the month my own website started paying me again: specificity is what turned recovery into growth, and it's the same specificity AI tools are hungry for.
• Step 7: Check back monthly, not weekly. AI indexing of new content is slower and less predictable than Google's. Expect to wait six to ten weeks before seeing any movement, and don't panic if week three shows nothing.

What gets picked up (with real examples)

Specific formats do better than others, consistently, in my own testing across client sites and my own.

• A named, dated case study beats a generic testimonial every time. "In March 2025, I helped a Bristol based bakery cut their Facebook ad spend from 800 to 300 pounds a month while keeping the same order volume, by moving budget into email" gets quoted. "Great results, would recommend" does not.
• A clearly answered FAQ question beats a paragraph buried in prose. If your page has a question as a heading and a direct answer underneath, AI tools can lift that pair almost intact.
• A number with context beats a number alone. "5 percent" means nothing on its own. "Google indexed 5 percent of 569 posts I published in 45 days" is a full, quotable fact, which is exactly why that post of mine keeps getting pulled into answers about AI content strategy.
• Original data beats repeated advice. If you've run a small survey of your own 40 customers, or you've got three years of your own pricing history, that's worth more to an AI tool than another summary of "five tips for social media marketing."

One small, easy win I recommend to almost every client: if you have any kind of interactive tool on your site, a pricing estimator, a quiz, a calculator, make sure the page around it explains in plain text what the tool does and what range of results people typically get. I wrote more on why interactive calculators convert so well, and the same logic applies here: an AI tool can't run your calculator, but it can quote the surrounding text describing what it does, so that text needs to carry real information, not just "try our free calculator."

What doesn't work, even though it's what most guides tell you to do

Adding an FAQ schema markup tag without changing the actual words on the page does very little on its own. The markup helps machines parse the structure, but if the answer underneath is still vague, there's nothing worth parsing.

Publishing more content faster is not the fix either, and I say that having tried it myself and written honestly about the result. Volume without substance gets ignored by both Google and AI tools. If you want the full, slightly painful numbers on that, I laid them out in the post about publishing 569 posts and getting 5 percent indexed. The lesson carries over directly to AI citation: quantity was never the answer, specificity was.

Chasing guest posts purely for backlink volume also does less here than people assume, though guest posting done, on relevant sites, with your own real expertise, still helps your overall authority. I wrote a full breakdown of how that works from both sides of the inbox in the editor's guide to getting a guest post accepted, and the same rule applies: editors and AI tools both reward specific, first-hand expertise over generic advice dressed up as a guest contribution.

A quick word on doing this yourself versus getting help

You don't need to be technical to run most of this audit. I'm not a developer and I do this on my own site every quarter with a spreadsheet and an hour of honest reading. I've written before about exactly how I use AI day to day without any coding background in how I use AI as a non-technical business owner, and the same low-tech approach works for this. Where it gets harder is knowing which pages matter most and how to prioritise a rewrite when you're already stretched running the business. That's usually the point where a short session with an AI implementation coach pays for itself, not because the work is complicated, but because an outside eye spots the vague pages you've stopped noticing.

How long this takes

Be realistic. I started rewriting my own case study pages with real numbers in early January 2026. The first AI referral traffic that could be traced to those specific rewrites showed up in week seven. It was small, 18 visits in that first week, but they were the right kind of visitor, people who'd asked a specific question and got sent to a specific answer on my site. By week twelve it had settled into a steady trickle rather than a spike, which is honestly the more useful outcome, because spikes fade and trickles compound.

If you want a lighter, faster win while the bigger rewrite is underway, spend twenty minutes building or improving one good resource list on your site. That kind of page gets referenced surprisingly often because it solves a specific need cleanly. It's the same reason a page like a list of 50 useful websites people don't know about keeps pulling in traffic long after it was published: it's specific, it's useful on its own, and there's something concrete to point to.

Related reading: 5 Social Media Tips For Small Businesses.

Frequently asked questions

Do I need to change my whole website to get mentioned by ChatGPT?

No. Start with your three most visited or most important pages, usually your homepage, your main service page, and one case study or about page, and rewrite the vague claims on those with real numbers and dates. Small, specific pages outperform an entire generic site.

Does this replace normal SEO?

No, you still need Google ranking for most of your traffic in 2026. Think of AI citation as an additional channel that rewards the same underlying habit, being specific and useful, but through a different mechanism than backlinks and keywords.

How do I know if my business is getting mentioned by AI tools already?

Check your website analytics for referral traffic from chat.openai.com, perplexity.ai, and Bing or Google AI features, then separately type your own name and services into each tool to see what comes back. Both checks matter because AI referral tracking is still inconsistent across analytics platforms.

What's the single fastest change I can make this week?

Rewrite one page, ideally your main service or about page, replacing every vague claim with a specific number, date, or named example. That one change does more for AI citation than any technical fix.

Related reading: The AI Subscription Stack: What I Pay For (and What I Cancelled) in 2026 and AI Phone Receptionists for Small Business: What They Cost and Where They Quietly Lose You Money.

I go much deeper on this in the AI marketing guide.