What AI uses to choose local businesses and what It ignores

If you sell digital marketing services to local businesses, you have probably spent the last year fielding questions about AI. Does it matter? Should clients be worried? What do they actually need to do?

We decided to stop speculating and find out.

Over the past several months, I conducted one of the most comprehensive studies into AI visibility for local businesses. I used the Insites platform to analyse 10,000 real businesses across ChatGPT and Perplexity, correlating over 200 digital marketing signals, including review volume, website content, technical SEO, social media presence, and structured data against real AI outcomes across three dimensions: awareness, reputation and recommendation.

My study found that 93% of the businesses in our study were known by at least one AI platform. On the surface, that sounds reassuring. But dig a little deeper, and only 55% had accurate information. The rest were being misrepresented with wrong addresses, incorrect opening hours and services that did not match reality. Even worse, AI was making things up entirely. For local businesses, it is an active problem that is costing them customers and potentially sending them to competitors instead.

Of all the signals we analysed, Google review volume showed the strongest correlation with AI visibility. Businesses found by both ChatGPT and Perplexity averaged 133 Google reviews. Businesses invisible to AI averaged just 11.

Google Business Profile completeness and Local Pack presence were equally significant, as businesses appearing in the Local Pack were recommended 3.5 times more often than those that were not.

It quickly became obvious to me that AI is drawing on the same trust signals that Google has relied on for years: social proof, verified presence, consistent information. Businesses that have neglected their GBP or deprioritised review generation are not just falling behind on Google, they’re becoming invisible to AI too.

But it’s not just about getting five-star reviews at volume. What those reviews actually say matters too. As search behaviour shifts towards conversational queries, people are no longer typing keywords like "plumber Manchester", they are asking AI "Which plumbers in Manchester can install an outdoor tap today at a good price?" The language inside reviews becomes a relevance signal. A review that mentions a specific service, a location and a timeframe gives AI exactly the kind of context it needs to make a confident recommendation.

Interestingly, a business's own website was the most frequently cited source across the entire study. ChatGPT referenced it in 72% of awareness and reputation queries. Even for recommendation queries, it appeared in over half of the cases.

But, AI reads a website differently from the way customers do. AI does not execute JavaScript or render pages in a browser. It reads raw content (everything in the code), not just what is visible on screen. Why does that matter? Well, it has some uncomfortable implications. AI can see your old copy, outdated service descriptions and leftover content that was never properly removed, which could negatively or inaccurately shape how a business gets understood.

What actually moves the needle on website visibility is depth and freshness. Businesses visible to AI had more than double the number of pages on average. Sites that had not been updated recently were significantly less likely to appear. But the goal is not a polished homepage and job done… Local businesses should be optimising with broad, specific, regularly updated content that clearly communicates what they do, where they operate, and who they serve.

This is the finding that will surprise most people in our industry.

Core Web Vitals, page speed, structured data, LLMs.txt, FAQ schema… None of these showed a meaningful correlation with AI visibility in our data. The technical tactics that have defined SEO best practice for years appear to have little bearing on whether AI surfaces and recommends a local business.

That does not mean a good website structure does not matter, but the signals AI responds to are fundamentally different from what the industry has spent years optimising for.

If you've been struggling to have conversations with local businesses about their AI visibility, this report makes things really simple. It's the biggest new sales and growth narrative for anyone selling to SMBs in over a decade because the conversation can completely shift. The local businesses that take it seriously now will be the ones winning as more and more customers use AI to search for recommendations and find out specific information about a business. Simply put, the businesses that will win in an AI-first world are the ones with the most detailed reviews, more specific website content, a complete and accurate GBP, and a consistent identity across every platform AI might look at.

For anyone selling to local businesses, that is both a challenge and an opportunity. The services that drive AI visibility are things you already offer. AI gives you a brand new and genuinely compelling reason to sell them.

The window to be early on this conversation will not stay open for long. Get in there now before your competitors beat you to it!