A homeowner in Austin asks ChatGPT, "who is the best plumber in town?" The assistant returns three businesses by name. Where did those names come from?
New research from Foundation Inc. on AI-mediated local discovery has the cleanest answer yet: overwhelmingly, Yelp. The pattern across ChatGPT, Google's Gemini, Perplexity, and Google AI Mode is consistent enough that it has stopped looking like a trend and started looking like a structural fact.
The Q4 2025 numbers
Foundation's research tallied 512,680 Yelp citations across the four major AI assistants in the fourth quarter of 2025. That is 3.4× more than the next platform on the list (Better Business Bureau, at 149,700). Between September and November alone, Yelp's citation share grew 19×, while overall AI citation volume grew roughly 3×. The trajectory is steeper than the market's, not flatter.
The breakdown by platform is more lopsided than the headline number suggests. On Google AI Mode, Yelp captured 72.5% of all local-business citations. On Perplexity, 62.1%. These are dominant shares, not competitive ones.
Why category queries matter more than brand queries
Inside the data, 91 to 97 percent of Yelp's AI citations originated from category queries ("best plumber in Austin," "emergency HVAC near me") rather than brand searches ("XYZ Plumbing reviews"). When a homeowner is in problem-solving mode — which is when contractors actually get hired — the AI is answering by combining a category and a location, then citing Yelp as the trust source for the business it names.
This rewrites what an "optimized" contractor digital presence looks like. The old model was: rank for thousands of long-tail keywords on a contractor website. The new model: dominate roughly 10 to 75 high-intent category queries that AI assistants repeatedly reference — and a meaningful share of that dominance now runs through Yelp.
The "near me" rewrite
The single most striking number in the research: of all "near me" citations across the AI assistants, more than 290,000 came from Yelp — roughly half of the total. The remainder, split across every other competing platform combined, contributed only about 87,000.
"Near me" was historically Google's high-intent local-search jewel. It has been quietly re-routed through Yelp by the AI layer. For a plumbing or HVAC contractor, that means "best plumber near me" is no longer answered by a Google Maps 3-pack the homeowner scrolls through. It is increasingly answered by an AI that consults Yelp and names three businesses. If yours is not one of them, you do not get the call.
The plumbing-specific multiplier
One statistic is hard to look away from if you sell plumbing services: in the Foundation data, 24 unique plumbing queries generated 1,168 Yelp-cited AI answers. That is roughly 49 cited answers per query — a multiplier that comes from the AI repeating Yelp citations across variations of the same underlying intent ("24-hour plumber Austin," "burst pipe repair downtown Austin," "best emergency plumber 78704," and so on).
Translation for a plumbing contractor: a clean, complete, well-reviewed Yelp business page is doing the work of hundreds of long-tail keyword variations that you would otherwise need to optimize for one at a time. That is unusually high leverage for a single surface.
Why this does not reverse quickly
AI assistants do not cite random websites. They need credible, verifiable signals that ground a recommendation in something the homeowner already trusts. Yelp supplies that scaffolding at scale: structured business profiles, photos, verified hours, response-rate signals, and a deep, time-anchored base of reviews tied to a single business entity.
When an assistant is constructing an answer, the path of least resistance is to cite the platform that has already done the verification work. As AI assistants compete on answer quality, dependence on trust-anchor platforms goes up, not down. Yelp's lead is structural.
What a contractor should actually do
- Claim and fully complete the Yelp business page. Category accuracy, service area, business hours, photos, response rate. An empty or stale Yelp profile is, at this point, a hole in your AEO infrastructure — not a missed nice-to-have.
- Treat Yelp categories as a strategic decision, not a checkbox. They are the connective tissue AI assistants use to map "plumber in Austin" to specific businesses. Specificity matters; generic categories underperform.
- Prioritize review velocity on Yelp itself, not only Google Business Profile. The AI assistants are weighting Yelp review depth and recency heavily in the Foundation data.
- Cross-link Yelp from the contractor's organizational schema via sameAs. This is how AI parsers stitch the Yelp signal back to the contractor's website and treat them as a single entity.
- Do not abandon Google Business Profile or the wider citation moat. But recognize that Yelp has moved, in the last six months of 2025, from one citation source among many to the dominant trust anchor for AI local search.
How TradeCraft Builds treats Yelp now
Yelp is one of the seven AEO surfaces TradeCraft Builds engineers as part of the standard 10-day sprint. The work is not just creating a profile. It is treating Yelp as a primary entity citation node — linked via sameAs from the contractor's Organization schema, monitored as part of the monthly citation-rate report, and optimized for the specific category queries that drive the contractor's market.
The Foundation research is the cleanest external confirmation yet of the discipline TradeCraft Builds was built around: engineering a contractor into the short list of businesses AI assistants name. That short list is no longer a theoretical artifact. It is the lookup table the assistants are increasingly using. The contractors who treat their Yelp presence accordingly in 2026 will compound a citation advantage through 2027 and beyond. The ones who do not will be invisible at the exact moment a homeowner is asking for help.