Ask ChatGPT "who should I call to replace a tile roof in Mesa?" and it will give you names, usually a handful, with a sentence about each. Those names aren't random, and they aren't ads. They come out of a retrieval process you can understand and influence.
How does ChatGPT generate a contractor recommendation?
For a local buying question, modern assistants don't answer from memory. They run a web search behind the scenes, pull a handful of pages they trust, and then write an answer grounded in what those pages say. That means every recommendation is really two decisions: which pages get retrieved, and which companies those pages support. You can lose at either step, either invisible to the search or present in the sources but not compelling enough to be named.
What sources do AI engines cite for trades questions?
The public citation studies agree on the broad shape. Yext's October 2025 analysis of 6.8 million citations across 1.6 million AI responses found 48.73% of ChatGPT's citations point at third-party sites like Yelp, TripAdvisor, and MapQuest, and for subjective "best X near me" queries, directory sources rise to 46.3% of ChatGPT's citations. The sources cluster into three buckets:
- Review and directory platforms: Yelp, Google, Angi, BBB, HomeAdvisor. In Local Dominator's analysis of 267,280 AI citations from local-business campaigns, Yelp led with 71,512 citations, ahead of Google at 55,977. The engines lean on these for trust signals like ratings, review counts, years in business, and license status.
- Community and social sources: Reddit (36,678 citations in the same analysis), Facebook, YouTube. Reddit threads read as third-party endorsements, and Perplexity in particular is built on them.
- Business websites: the smaller bucket by volume, but the one you fully control, and where engines look for prices, response times, and service specifics no directory publishes.
What gets a company from cited to recommended?
Getting your pages retrieved is necessary but not sufficient, because the engine still decides which companies to put in the answer. Watching thousands of answers, we see the named companies share a recognizable profile:
- A strong, recent review record. Rating and volume both matter, and engines notice recency: a 4.7 with fresh reviews beats a 4.9 that went quiet in 2024.
- Consistent identity everywhere. Same name, phone, service area, and trade across your site, Google Business Profile, and directories. Mismatches read as uncertainty, and engines skip uncertain candidates.
- Verifiable specifics: license numbers, years operating, real project photos, named neighborhoods. Engines prefer claims they can ground in a source.
- Content that answers the question asked. If the question is about cost, pages with real numbers win. If it's "emergency repair", pages that state a response time win.
Do different AI engines recommend different contractors?
The same company can be a fixture on Gemini and nearly absent from ChatGPT, because each engine builds its answers from different sources. Ahrefs found 88% of ChatGPT's cited URLs come from its general web-search retrieval rather than news, Reddit, YouTube, or academic sources, and Profound found Reddit made up about 46.7% of Perplexity's top cited domains on commercial queries in 2025, a share that had fallen to roughly 24% by January 2026, while Gemini leans on Google's local graph (Maps, reviews, Business Profiles). So you can't check one engine and call it done. Your review profile might carry you on one while a missing cost guide sinks you on the other.
What should a contractor do with all this?
- Audit the question set, not the engine. List the 20–50 buying questions that matter in your market. That list, not any single answer, is the battleground.
- Publish the pages the engines are starving for: a priced cost guide per major service, a real page per city you serve.
- Tighten the trust layer: review velocity, consistent NAP data, license info on your site.
- Track per engine, weekly, and let the deltas tell you what's working.
