Introducing the Insites Local AEO Benchmark: monthly data on how AI recommends local businesses

Today we're launching the Insites Local AEO Benchmark, a free, monthly measure of how AI assistants identify, describe and recommend local businesses.

It covers seven of them: ChatGPT, Gemini, Copilot, Perplexity, Grok, Google AI Overviews and Google AI Mode. The first edition is built on a sample of 8,016 local businesses audited in September 2026, and it'll update every month from here on.

Local AI visibility is a different problem

Most AI visibility tools and AEO studies are built around brand awareness. They ask whether ChatGPT mentions a national brand, how often, and in what tone. That's a fair question for a trainer brand or a bank. It's the wrong one for a plumber, a dentist or a family-run restaurant.

Local businesses don't win or lose in AI on brand recognition. They win or lose on one question: when someone asks for the best emergency plumber in their town, does AI put their name forward? That answer depends on the location, the service and the competition down the road, and it changes from town to town and trade to trade. You can't measure it with generic prompts, and you can't learn much about it from a few dozen businesses.

Why we're the ones publishing it

This is where we're in an unusual position.

• We look at small businesses, at scale. Every month, Insites customers run AI visibility audits on thousands of real local businesses across many trades and towns. That's the sample behind the Benchmark, and it grows every month.
• We ask location-specific questions. Each audit asks the assistants what a customer would actually ask, naming a trade and a town, and records which local businesses get recommended and in what order. It's not a brand-mention count.
• We have the rest of the picture too. For most of those businesses we also hold the traditional signals: Google Business Profile, reviews, local rankings, website health and SEO. That means we can go beyond what AI says and look at why, by correlating local AI visibility against the things agencies already work on. It's how our AI Visibility Report was able to show which signals separate the businesses AI recommends from the ones it ignores.

Put together, it's a view of small businesses in AI search that nobody else has.

Why we're opening it up

We could keep this data inside the platform. We're publishing it instead, free, for anyone to use, because we think the market badly needs it.

There's a real shortage of good information about AI and local businesses. A lot of the statistics doing the rounds come from companies selling the fix, and the numbers tend to be chosen to scare rather than inform. "AI can't see your business" makes a great headline for a tool that promises to fix it. It's also, for most local businesses, not true: in our first edition, AI found 99.1% of the businesses we checked.

A lot of agencies, resellers and media sales teams are now selling AI visibility to SMBs, many of them to hundreds or thousands of clients at once. We want that work to be good. A market full of scare stories and quick fixes that don't move anything is bad for SMBs, bad for the people selling to them, and bad for us.

Delivering a great AI visibility service starts with understanding the market: what AI actually does with local businesses, what moves a business from known to recommended, and what doesn't. That's what the Benchmark is for. It has three jobs:

• Show what's really happening. Real local businesses, location-specific questions, seven assistants, measured the same way every month.
• Separate what works from what doesn't. So you spend your clients' budget on the things that change whether they get recommended.
• Give you a figure you can defend. Every number carries its month and sample size, so you know exactly what you're quoting.

Who it's for

The Benchmark is for anyone who sells or delivers digital marketing to local businesses: agencies, resellers, telcos, directories, media companies and the sales and fulfilment teams inside them. It's especially useful if you're delivering AI visibility across a large base of SMB clients and need to know where to focus.

It assumes you already know what AEO and AI visibility are. It isn't trying to convince you the channel exists. It's trying to help you understand it, explain it to a client, and turn it into work that gets results.

It's not written for SMB owners directly, although you're very welcome to show it to them. Each finding comes with a line written for an owner rather than a marketer.

What the first edition found

The headline is one we think every agency should be leading with: being known by AI and being recommended by AI are two very different things.

• AI found 99.1% of the local businesses we checked, across at least one of the assistants we asked.
• But 40.3% of those businesses were never recommended for a single one of their services when a customer asked for one.
• ChatGPT knew the business by name 96.0% of the time, and recommended it only 27.8% of the time.
• Where a business publishes its phone number, AI got it right only 53.2% of the time. 94.3% of businesses had their name, website or phone number wrong in at least one assistant's answer.
• Only 13.2% of the sources AI cited about a business were the business's own website. Most of what AI says about a local business comes from sites it doesn't run.

That first point matters for how you sell. Telling a local business owner they're invisible to AI is one prompt away from being disproved in front of you. Telling them AI knows who they are but sends customers to a competitor when someone asks for what they do is true for most of them, and it's a much more useful conversation.

The Benchmark also breaks the data down by assistant, by sector (interior design and flooring businesses had the hardest time getting recommended this month, security businesses the easiest), by what AI says about businesses when it's less positive, and by the website readiness checks most sites fail.

On that last one, a word of caution. The readiness checks, such as FAQ and LocalBusiness schema or an llms.txt file, are cheap, sensible hygiene. They're not what gets a business recommended. Our own research found reviews, complete listings and a genuinely useful website do far more of the heavy lifting, and review volume in particular is the strongest signal we've found.

How to use it

We built the Benchmark to be picked up and used, not just read.

• Quote it freely. Everything on the page is free to use in your own content, proposals and pitch decks. Just say which month the figure is from and how many businesses it covers, and link back to the Benchmark if you can.
• Copy the client-ready lines. Each finding has a short "How to communicate this to SMB owners" paragraph you can lift straight into an email or a proposal.
• Present it. The Present button opens a full-screen, client-safe view with one finding per slide, and there's a printable handout version too.
• Switch the timeframe. Read the latest month, or the last three months for a steadier view.

How it's measured

Each audit puts a business to up to seven AI assistants and asks two kinds of question. A branded question names the business and checks whether AI can find it and get its details right. An unbranded question asks what a customer would ask, such as the best plumber in a given town, and records which businesses get recommended and in what order.

We're open about the limits. The sample comes from businesses our customers chose to audit, not every local business, so read the figures as direction rather than precision. AI assistants are non-deterministic, so small moves are marked steady rather than reported as change. The full method sits at the bottom of the Benchmark page.

What happens next

The Benchmark updates early every month with the month just gone. If you'd like to know when it lands, The Deep Dive is our monthly newsletter that goes with it, with forecasts, extra data that doesn't make the page, and ideas for putting it in front of clients.

And if you want the same picture for one specific client rather than the market as a whole, that's what Insites does. Run a free AI visibility check on any local business, or see how AI Visibility Monitoring tracks it for your whole client base.

Explore the Local AEO Benchmark