Why topping Google means nothing to ChatGPT right now

New research shows that being the best-known or best-performing business in your market has almost no bearing on whether AI assistants recommend you. A study of 37,500 AI searches found that 91.5% of established, top-producing real estate agents were never named once by any AI platform, and a separate study of AI answers to everyday questions found that page-one Google rankings made almost no difference to who got recommended. The common thread: AI systems reward businesses that publish citable content about themselves, not businesses that simply perform well or rank well.

What changed?

Local search analytics firm Local Falcon published its Realtor AI Visibility Index in August, and the findings are still rippling through the industry this week. The firm ran 37,500 AI searches across the 100 largest US cities to find out which real estate agents AI names. The results were stark: among the ranked, website-publishing agents and teams checked in this study, 91.5% were never cited once.

Being good at the job didn't help either. Only 21.2% of the highest-volume individual agents in these cities were cited by any AI platform. And there's no single platform to win over, since more than eight in ten agents were named by only one platform and by no other.

This week, marketing firm Top Real Estate Systems launched a free AI Visibility Report built directly on the Local Falcon data, pushing the findings further into mainstream attention. The report shows agents their Google map ranking alongside whether ChatGPT, Gemini, Google AI Mode, AI Overviews and Grok name them.

Local Falcon's CEO put the disconnect bluntly. As he told Inman News, an agent can be number one in their city by sales volume and still not exist as far as ChatGPT is concerned, because the models aren't reading MLS data or production rankings, they're reading pages.

The pattern isn't unique to real estate. A separate study reported by Dynamic Business this week, from Sydney agency EarnedMedia, looked at 360 responses from ChatGPT, Gemini, Claude and Perplexity to 90 everyday Australian questions. It found that businesses in the top 25% for mentions and links across other websites were recommended 1.9 times as often as those in the bottom 25%. Crucially, ranking on page one of Google made almost no measurable difference, and company size and marketing budget didn't change the result.

A third data point adds a wrinkle: AI recommendations aren't even stable from one query to the next. In a June 2026 study by Dossis Digital, ChatGPT kept the same recommended businesses across three identical runs just 8% of the time, while Gemini managed 24%, Perplexity 38% and Claude 89%.

Why does this matter for your business?

If you've assumed that strong reviews, market share or a high Google ranking will naturally carry over into AI visibility, this data says otherwise. The signals AI models actually use are different: who writes about you, how often you're mentioned on other sites, and whether your own content exists in a form the model can lift and cite.

This creates a strange incentive. Local Falcon found that the agents who did get cited often weren't the market leaders, they were the ones who'd published their own "best agents in [city]" style content. As the CEO put it, the agent who wrote a "best agent" post gets named, and the agent who actually sells the most homes doesn't have that advantage by default.

For founders and marketing leaders, the takeaway is the same across industries: production, revenue and reputation are invisible to AI unless they're expressed somewhere as text the model can find, trust and quote. A business can be the undisputed leader in its market and still be structurally absent from the answers customers are now asking for.

The instability finding matters too. If an AI assistant recommends different businesses 60-90% of the time it's asked the same question, that's not just a visibility gap, it's a volatility problem. A business might appear in an answer today and disappear tomorrow, with no change in its underlying quality or service.

What should you do now?

Start by finding out where you currently stand. Ask ChatGPT, Gemini, Perplexity and Google's AI Overviews the exact questions your customers would ask, and see whether you're named at all, and if so, by which platforms. Given the Dossis Digital findings on inconsistency, run the same question more than once before drawing conclusions.

Second, look at who is getting cited instead of you, and work out why. Local Falcon's research found that much of the time, AI answers trace back to agents who publish a website, where AI cited 8.5% of them, a small but telling slice compared with the broader pool. Winning agents weren't necessarily better performers, they'd simply created content positioning themselves as an authority in a format AI could read and repeat.

Third, treat earned mentions and third-party citations as seriously as your own site content. The EarnedMedia research suggests that being talked about elsewhere, on other sites, in other people's content, matters more to AI recommendation odds than where you sit in Google's organic results. That means PR, guest content and structured data about your business matter more for AI visibility than classic on-page SEO tactics alone.

Finally, don't treat a single AI visibility check as the final word. If you want an ongoing view of how your brand shows up across ChatGPT, Gemini, Perplexity and AI Overviews, it's worth running a free audit, such as the one from Sited, to see where the gaps are before a competitor fills them with their own "best of" content.

Frequently asked questions

Does ranking well on Google guarantee I'll be recommended by AI tools?

No. The EarnedMedia study found that ranking on page one of Google made almost no measurable difference to whether a business was recommended by AI assistants answering everyday questions.

Why do top-performing businesses get left out of AI answers?

Because AI models work from published text, not internal performance data. Local Falcon's research found that the models aren't reading MLS data or production rankings, they're reading pages, so a business has to describe its own strengths in content the model can find.

How consistent are AI recommendations if I do get cited once?

Not very, and it varies sharply by platform. The Dossis Digital study found ChatGPT kept the same recommended businesses across three identical runs just 8% of the time, while Gemini managed 24%, Perplexity 38% and Claude 89%, so one good result doesn't mean you'll show up again next time.

What's the single biggest factor in getting cited by AI search tools, based on this research?

Third-party mentions and links appear to matter more than almost anything else measured so far. Businesses in the top 25% for mentions and links across other websites were recommended 1.9 times as often as those in the bottom 25%, regardless of company size or budget.

Does this only apply to local businesses like real estate agents?

No. While the Local Falcon study focused on real estate, the EarnedMedia study covered general consumer questions across industries and found the same pattern: citation-worthy content and external mentions outweigh traditional ranking signals in AI-generated answers.