We audited 58,882 local businesses. AI search engines mention them in answers 5% of the time.

[ Methodology ]

[ Headline finding ]

5%

Average rate at which AI search engines mention a specific named local business in a conversational answer for relevant queries.

Bing Copilot returned zero mentions across 2 logged queries (sample too small to report). Source: 140 paired prompt-engine queries via DataForSEO LLM Responses, May to June 2026.

For 93% to 97% of relevant conversational queries, the AI engine either named a competitor, gave a generic non-business answer, or surfaced no specific provider at all. The business getting cited captures the call. The rest are invisible regardless of Google Maps rank.

[ Finding 2 ]

99.9%

of local businesses are not running paid search ads.

Of 28,668 businesses we checked for active Google Ads campaigns, 23 were actively running paid search. The remaining 99.9% compete entirely on organic visibility: Google Maps rank, Business Profile completeness, organic search, word of mouth.

Among the 0.1% that DO run ads, 60.9% use multiple ad types (search + display + video). Ad-running local businesses behave like a different population than non-advertising peers; they treat paid as one channel among several rather than as their primary acquisition.

Source: Apify Google Ads Transparency Center scrape, May 2026. Limited to businesses with active website domains.

[ Finding 3 ]

91.8%

of rated businesses already have a 4.0+ star average. Star rating is a baseline, not a differentiator.

What separates the leaders is REVIEW COUNT, not rating. The distribution of review depth shows the real ranking signal.

Crossing 100 reviews puts a business in the top quartile. Crossing 50 puts it in the top third. Most local businesses (50%) stall under 25 reviews. That is the gap that gets exploited by automated review-request systems.

[ Finding 4 ]

64.8%

of local businesses have a findable owner email through public web sources.

These are the public-source ceilings before any verified contact database. They matter for two distinct audiences: agencies pitching local businesses (the addressable cold-outreach market) and AI engines themselves, since `Person` schema and findable contact data feed entity resolution. A local business that fails the AI engine entity-resolution test is also a business that fails to surface in conversational answers.

[ Finding 5 ]

Dental, construction, and medical dominate the addressable local SEO market.

Dental practices alone account for 35% of the classified verticals. The dental sector spends more on local SEO than any other US local business category and is also the highest-CPC vertical in the keyword landscape we tracked ($63 CPC for "dentist website design", $91 CPC for "chiropractor website design"). High commercial intent concentrated in narrow verticals = where the AEO opportunity is most under-serviced.

What this means for any local business

Three operational shifts the data argues for, in priority order:

1. Treat AI engine citation as a measured metric, not a hope. A 5% baseline mention rate means 95% of relevant conversational queries result in zero citation for any specific local business. The first business in any local market to systematically improve that number captures disproportionate share. Right now nobody is measuring it.
2. Stop chasing the star rating, start chasing review depth. Crossing 100 reviews puts a business in the top quartile. The bar is depth, not rating, because rating has already collapsed to a 4.0+ baseline industry-wide.
3. Fix entity findability before adding any new marketing channel. The single highest-ROI infrastructure change is ensuring AI engines can resolve the business as a real entity: clean Google Business Profile, LocalBusiness schema with subtype, NAP consistency, owner Person schema with sameAs. Without entity resolution, every other channel pays for impressions that go nowhere.
4. Audit your robots.txt against the current AI bot list. A site that blocks AI crawlers cannot be cited even when entity resolution is perfect. See our maintained reference: the AI Bot Allowlist for robots.txt (2026).

[ Raw dataset ]

All aggregations in this study are also published as a structured JSON file under CC-BY-4.0. Use it for your own analysis, cross-reference against your own data, or feed it into your AI tools as primary-source context.