How do contractors show up in ChatGPT and AI search?

July 27, 2026, updated September 8, 2026 · Joshua Hill

A growing share of homeowners has stopped typing "painter near me" into Google and started asking an assistant a full question: who should I hire to repaint my house in Bellingham, and what should it cost? The assistant answers with two or three company names. Either you are one of them or you are invisible.

What the AI tools actually read

There is no secret index for AI search. ChatGPT, Google's AI results, Perplexity, and the rest assemble answers from the same public web you already live on: your website, your Google Business Profile, your reviews, directories, and anywhere your company gets mentioned. The difference is how they read it. A search engine matches keywords and shows a list. An assistant reads pages, extracts facts, and composes an answer. That changes what wins.

Seven things that get a contractor into the answer

• Let the search crawlers in. Many sites block AI crawlers without knowing it. Check your robots.txt for the crawlers that feed answers: OAI-SearchBot is what ChatGPT search uses, PerplexityBot feeds Perplexity, and ordinary Googlebot is all Google needs for AI Overviews and AI Mode. Those are separate from the training crawlers. OpenAI's GPTBot and Google's Google-Extended govern whether your pages may be used to train models, and you can block those and still be eligible for search answers. Allowing training is a choice, not a price of admission. If the search crawlers cannot read you, they cannot recommend you.
• Answer questions in plain sentences. Assistants quote pages that state facts directly. A page that says what you do, where you work, and what a typical job involves in ordinary prose beats a page of slogans every time.
• Keep your facts identical everywhere. Same business name, same city, same phone, same services on your site, your Google Business Profile, and every directory. Assistants cross-check sources, and inconsistency reads as doubt.
• Use structured data. Schema markup for a local business, your services, and your articles hands the machines your facts in their native format instead of making them guess.
• Consider an llms.txt file, and know what it is. A plain-language summary of your business at yourdomain.com/llms.txt is a proposed convention, not a standard any major assistant has said it uses for ranking. Google has said its AI features need nothing beyond normal search eligibility: no special file, no special schema. We publish one because it costs ten minutes and does no harm. Do not pay anyone who calls it a ranking secret.
• Earn reviews that mention specifics. A review that names the town, the trade, and the job gives an assistant exactly the evidence it needs to match you to a question.
• Keep publishing. Fresh, useful pages about real questions in your trade give assistants more surface to find and more reasons to treat your site as a source worth citing.

What nobody can promise

Nobody can guarantee placement in an AI answer, and anyone who promises it is selling something they do not control. What you control is whether the machines can read you, whether your facts are consistent and specific, and whether there is enough real substance on your site to be worth citing. Do those things and you are ahead of most of your market, because most of your market has not started. Whether any of it is working is a separate question, and how to tell if ChatGPT is sending you customers covers what the tracking can and cannot show.

Where this fits

AI search visibility is part of the Get Found layer of the pipeline we build, alongside local search and the map pack. It is the same discipline applied to a new reader, sitting on top of the page-level groundwork in what a contractor website actually needs. We wire it into our own companies' sites first, which is also how we know which parts matter and which parts are noise.