680 AI Citations, 1 Visit: Is AI Visibility Worth It for Small Businesses?

Last week I opened the AI Performance report in Bing Webmaster Tools and saw something that made me laugh and frown at the same time. One of my blog articles had been cited by AI 680 times. The number of visits it brought? One.

The article was my step-by-step plan for optimising a website for ChatGPT and AI search. When I looked at the queries AI used to find it, I noticed it was being cited for something it didn't really answer – whether small businesses should pay for AI tools. So I immediately wrote a new article that does: Is it worth paying for ChatGPT, Claude or Gemini if you run a small business? Two people in the LinkedIn comments later called this a free content report delivered to me by AI – and I couldn't agree more with them.

vs.

I shared it on LinkedIn, and it turned into one of the liveliest discussions I've had – SEO specialists, AI search consultants and fellow marketers all jumped in. Some agreed, some corrected me, and a few made me rethink how I read the data. In the same week, three new studies on AI traffic came out, and Rand Fishkin of SparkToro summed them up in a way that made a lot of marketers pause.

This article pulls it all together – what AI citations really tell you, where the clicks go, and how to report AI visibility to a CFO without a 20-tab dashboard.

What do AI citations actually tell you?

In February 2026, Bing added an AI Performance report to Webmaster Tools. It shows how often your pages are cited in Microsoft Copilot, Bing's AI summaries and some partner tools, plus "grounding queries" – the phrases the AI used to find content it then cited in its answers.

Here's what I learned, thanks to the comments under my post:

• Grounding queries aren't what people typed. They're search phrases the AI creates itself to find sources for its answer.

• A citation isn't a person. 680 citations doesn't mean 680 people saw my link.

• Bing shows only Microsoft's corner of AI – not ChatGPT, Claude or Google's AI Overviews.

So the data is directional: it shows you a trend, not an exact count. Still useful – just not something to put in a report as hard fact. Read grounding queries as the topics AI is researching, and they become one of the best free sources of content ideas you'll find.

How much traffic does AI really send?

Not much – and now there's solid data behind that.

Marie Haynes analysed 12 months of Google Analytics data from 100 websites – over 145 million sessions. All AI assistants combined sent less than 1% of the traffic Google organic search sent. ChatGPT visitors did convert better (13.8% vs about 3% for Google organic), but Google still delivered 96.5% of all conversions.

The SEO Works tracked 95 websites for a year and found AI assistants sent about 2.6% as much traffic as organic search. AI visitors converted only slightly better on average – and on more than a third of the sites, they converted worse. For every 100 conversions from organic search, AI brought about 3.

After reading the studies, Rand Fishkin said he felt less confident than ever that visibility in tools like ChatGPT has much business impact for most companies – while adding that Google's AI Overviews and AI Mode are a different story.

Is AI traffic hiding in "Direct"?

This was my favourite theory – and one I repeated in the LinkedIn thread. Many AI apps don't pass on where a visitor came from, so Google Analytics files those visits as "Direct". So maybe my one visit was really ten?

Kevin Indig tested exactly that across 12 websites. The Direct traffic spikes he found were mostly bots, and visits with traces of AI were a tiny fraction of Direct traffic. He also checked whether AI visibility leads to more branded searches – and the early links he found didn't hold up under a stricter test. The SEO Works study came to a similar conclusion: no meaningful link between AI growth and Direct traffic growth.

So, honestly? My one visit might really be about one visit.

Is AI search a catalogue – or a distraction?

My theory in the LinkedIn thread was that AI search works like the Lidl catalogue. You flip through it, you know the products exist, but you don't buy from the catalogue – you go to the store.

The studies don't prove that theory. But they can't disprove it either: analytics simply can't see someone who read about you in ChatGPT and came back three weeks later through a Google search. That's an awareness effect, and awareness has always been hard to measure – ask anyone who defended a print or TV budget when digital took off.

What I take from it for small and medium-sized businesses:

• Don't treat AI citations as a goal on their own. They're a side effect of good content, not a KPI.

• Keep Google first. It still sends the vast majority of visitors and conversions – and its AI Overviews matter.

• Write content that answers real questions clearly. It serves Google and AI at the same time.

How do you report AI visibility to a CFO?

Imagine your CFO says: "AI citations bring no leads, so let's stop investing in AI search." What do you show? Here's the simple monthly snapshot I'd use, from the strongest signal to the weakest:

1. How customers say they found you. Add a "How did you hear about us?" field to your forms and ask on sales calls, with ChatGPT/AI as one of the options. Kevin Indig sees this kind of self-reported answer as the more useful signal – and it's free.

2. AI visits and what they do. Google Analytics now has an AI Assistant channel. Report visits and conversions – next to your total traffic, so the numbers stay in proportion.

3. Whether you show up. Once a month, ask ChatGPT and Google's AI 10–20 questions your buyers would ask. Are you mentioned? Note yes or no.

4. Branded searches. Check the trend in Google Search Console – as supporting evidence, not proof.

No single number makes the case. But if customers start mentioning AI and you keep appearing in the answers, that's a story a CFO can follow.

So, is AI visibility worth it?

Nobody has a perfect answer yet – not the tools, not the studies, not the experts in my comments. What matters is reading the data honestly and spending where the evidence is. And as I wrote in my post: at the end of the day, you write content for people. If they don't find it useful, they won't stay – even if AI recommends it.

How do you measure AI visibility? I'd love to hear what's working for you.

Not sure whether your marketing is building visibility or quietly leaking budget? A marketing audit shows you what's actually working – and what isn't.

I'm Tereza – I help founders stop being the marketing department and start having one.

FAQ

Sources

• Marie Haynes, GEO is a distraction (October 2026)

• The SEO Works, AI referral traffic 2026: how much, and does it convert? (October 2026)

• Kevin Indig, Growth Memo, Does AI drive more direct traffic? (October 2026)

• Rand Fishkin, LinkedIn post on the three studies (22 October 2026)

• Practical Ecommerce, Bing adds AI visibility reporting

• Momentic, Initial thoughts on AI Performance reports in Bing Webmaster Tools

• Themeisle, How to track AI referral traffic in GA4