AI assistants concentrate local business recommendations in three names

New data from MentionedOn shows AI assistants heavily concentrate local-service recommendations, with the top three businesses taking 56.2% of all mentions across 840 leaderboards in 40 U.S. metros. The pattern matters because it suggests that ranking first in AI-generated local answers may be far more valuable than simply being listed.

Why it matters: - AI assistants are not surfacing broad local directories. They are compressing answers into very short shortlists. - For a local business owner, placement near the top can matter far more than making the list at all. - The study suggests that visibility in AI local search is highly unequal by trade, with some categories dominated by chains and others still wide open for independents.

What happened: - MentionedOn analyzed 840 local-service leaderboards across 21 trades and 40 U.S. metropolitan areas. - The data covered the most recent leaderboard per trade and metro as of 2026-08-22. - AI assistants named an average of 9.2 businesses per trade and metro, with a median of 10. - The top three businesses in each trade-and-metro combination captured 56.2% of all mentions. - The single most-mentioned business captured 24.7% of mentions. - The dataset included 5,552 distinct businesses.

The details: - National brands accounted for 54 brands and 17.9% of all mentions. - MentionedOn defined a national brand as a name appearing in at least 10 of the 40 metros. - Two Men and a Truck and Terminix were the only brands named in all 40 metros. - Moving saw 51% of mentions go to brands appearing across at least a quarter of the metros studied. - Pest control saw 42% of mentions go to those multi-metro brands. - Accountants saw 37% of mentions go to those multi-metro brands. - In five of the 21 trades, chain dominance was effectively zero. - The five trades with almost entirely independent local results were HVAC, law, veterinary care, real estate and med spas. - MentionedOn generated each leaderboard by asking two AI assistants five category questions and extracting the business names they recommended. - The study used 8,400 assistant answers from a corpus that began on 2026-06-18. - The 21 trades and 40 metros were selected by MentionedOn and are not a random sample of U.S. local businesses. - Mention counts reflect recommendation frequency, not revenue, quality or customer satisfaction. - More information is available in the company's announcement.

Between the lines: - The core pattern is concentration, not blanket chain dominance. - Some trades appear highly standardized in AI answers, while others still leave room for independent businesses to win recommendations. - The study implies that businesses competing for AI visibility may need to think less about broad presence and more about cracking the top tier of recommendations. - John Arndt, founder of MentionedOn, said the more useful question is whether AI puts a business in the first three names, because the difference between being listed and being invisible gets very small after that.

What's next: - MentionedOn's leaderboard corpus will likely keep changing as assistants are updated and local markets shift. - Businesses in concentrated categories may need to monitor whether their names appear in the first few recommendations, not just whether they appear at all. - Independent operators in lower-concentration trades may still have room to gain share before national brands dominate the category.

The bottom line: - AI local recommendations are shaped less like a directory and more like a winner-take-most shortlist.