When you ask ChatGPT, Claude, or Perplexity to recommend a dentist in Portland, an ecommerce store for outdoor gear, or a mortgage broker who works with first-time buyers, the answer comes from a small group of sites the AI bots have decided are worth citing. We wanted to know which sites those are. So we cohorted the crawl data and looked.
What we found was not what we expected.
We analyzed 1,187,128 brands in the Engagemii directory. Of those, 395,022 had been crawled at least once by a major AI bot. We sorted that crawled population by total crawl volume and split it into cohorts. The top 0.1% (just 395 sites) averaged 38.8 AI bot crawls each. The bottom 90% averaged 1.3.
That gap on its own is interesting. A 30-times multiplier between the head and the tail of a long-tail distribution is not unusual on the web. What was unusual was who those 395 sites are.
Almost none of them are Fortune 500 companies. We checked. The top of the AI-crawl distribution is dominated by local dentists, regional ecommerce stores, niche professional services firms, mortgage brokers, specialty supplement brands, and the kind of small businesses that would never break a SEMrush top-100 list. There is a Fortune 500 presence in the data, but it does not sit at the top. The very top is held by businesses you have probably never heard of.
There is a simple answer and a more interesting one. The simple answer is that those 395 small businesses have done the technical work that makes them readable to AI engines, and most large enterprises have not. The Engagemii AEO score measures readiness across six categories: structured data, content structure, entity clarity, E-E-A-T credibility, technical infrastructure, and AI crawler access. Sites that score 7 or higher are consistently citable. Sites below 5 usually are not.
In the top 0.1% cohort, 29.9% of sites have an AEO score of 8 or higher. In the bottom 90% of crawled sites, only 0.4% do. That is a 75-times difference in score-8-or-better adoption between the head of the AI-crawl distribution and its tail.
The more interesting answer is that AI bots are an enormous reset button for who gets discovered online. Traditional Google rankings have a strong correlation with brand recognition: famous brands rank well partly because they are famous, partly because they have the link profiles and content investments that fame produces. AI bots have to work with the actual structured signals on the page. They cannot consult a Trust Pilot rating, a New York Times mention, or a billion-dollar marketing budget. They can only read what is in the HTML. If a brand has not made itself machine-readable, the brand does not exist in the AI engine's mental model. Size does not save it.
We pulled examples from our directory. A two-location dental practice in the southeast averages 47 AI bot crawls per month. Their site has been audited and tuned. JSON-LD Organization schema is present. FAQ schema covers the questions patients actually ask. The robots.txt explicitly allows GPTBot, ClaudeBot, PerplexityBot, and Google-Extended. The site has an llms.txt manifest at the root explaining who the practice is in plain English for AI agents.
A regional outdoor gear store with about 40 employees has similar numbers. Product schema is on every SKU. Author bylines run on every blog post with credentials in machine-readable form. Their AEO score is 8.
A mortgage broker in a single midwestern metro averages more AI crawls per month than the brand page for a Fortune 50 retailer. The retailer has more traffic, more backlinks, more brand recognition. The broker has better schema. AI bots reward schema.
None of these brands have a marketing budget that competes with the enterprise companies they outperform in AI citations. They have a technical advantage that costs roughly the same as updating a robots.txt file and adding a few JSON-LD blocks to a site template. They just decided to do it sooner than most.
The data tells you two things. First, AI search is a more level playing field than traditional Google search. Brand recognition matters less. Backlinks matter less. Structured data matters more. If you have been waiting for the channel where small businesses can outcompete enterprise budgets, this is it.
Second, the window is open right now, but it will not stay open forever. Eventually large enterprises will hire AEO specialists. Eventually their schema will be as clean as a thoughtful dentist's. But that has not happened yet. Most enterprise sites we audit score between 3 and 5. The same as the median small business. The only sites consistently in the 8-plus range right now are the ones run by owners who personally cared about being cited and went and did the work.
If you are a local business with 1 to 100 locations, this is the moment to invest in your AEO setup. The capital cost is low. The implementation can be done in a single afternoon for most sites. The competitive advantage compounds because once an AI bot sees you as a high-quality source, you are in the shortlist it returns to.
If you are an enterprise marketer reading this, the urgency is different but real. Your competitors are not the brands you have always benchmarked against. Your competitor in AI search is the small business in your category that has decided to take this seriously. They will appear in answers where you expect to. They will get cited in places you have not realized AI traffic is now being routed through.
Engagemii's AEO score is free. Enter a URL at engagemii.com/aeo and you will get an audit across all six signal categories in about 30 seconds. The score is a 0-to-10 number with a category-by-category breakdown showing exactly where your signals are weak. If the score is below 7, the report also returns a prioritized fix list ranked by impact on the underlying model that drove this research.
Most sites land between 3 and 5 on first audit. That is normal. The path from 5 to 8 is the part of the work that takes some effort, and the path from 3 to 5 is usually one afternoon. The point is to know where you sit before you decide how much effort to invest.
The numbers in this article come from Engagemii's open research dataset, generated 2026-05-29. The dataset contains structured-data feature flags, AEO scores, and observed AI bot crawl counts for 1,187,128 brand domains. The cohort analysis described here is Section 3.2 of the full research brief. The full methodology, including the two-stage LightGBM model used to attribute crawl behavior to specific signals, is at engagemii.com/research/aeo-crawl-drivers.
If you want to cite this article, the URL is engagemii.com/blog/ai-bots-cite-small-businesses-not-fortune-500. We publish a new finding from this dataset every week.
Ready to find out if AI can cite your brand?
Get Your Free AEO Score