How to Show Up When a Homeowner Asks ChatGPT for a Contractor in Central Oregon

Per SOCi data published in 2026, only 1.2 percent of local businesses get recommended by ChatGPT. ChatGPT now has 900 million weekly active users. The math is ugly: a homeowner in Bend asks ChatGPT for a roofer, and 98.8 percent of the time, your shop is invisible to the answer. Here is what actually controls whether you show up, and what to do this quarter.

How ChatGPT actually picks a contractor

The first thing to understand is that Google Business Profile, the thing every contractor has been told to optimize for the last decade, does not feed ChatGPT. ChatGPT pulls local business data primarily from Foursquare (about 70 percent of the local data signal) and layers Bing search results on top of that. When a homeowner asks "who is the best roofer in Bend, Oregon," ChatGPT runs a live Bing search, scans roughly the top 20 to 30 results, and re-ranks based on its own criteria. Per Yext's analysis of 6.8 million AI citations, only 45 percent of the brands that rank well on traditional Google search also show up in AI citations. Which means the work you have done for Google SEO transfers about half the time, and the rest is a different game.

Perplexity is a different beast. It leans web-first and pulls heavily from industry-specific directories. For contractors, that means Angi (which is now embedded inside ChatGPT through a documented integration), BBB, Yelp, and trade-specific aggregators. Google AI Overviews, by contrast, do pull from Google Business Profile plus organic top-10 plus reviews. Three different AI surfaces, three different mechanics. The contractor who treats them as one problem ends up invisible on at least two.

The structured data play

Here is where most contractor sites fail. They use generic LocalBusiness schema. The schema.org spec has specific subtypes that AI systems extract more reliably: RoofingContractor, HVACBusiness, Plumber, Electrician. Use the specific subtype. Per analysis from contractors who track AI visibility, the specific-subtype move alone produces measurable lift in citation frequency. JSON-LD format. Google prefers it, AI extractors handle it cleanly.

Beyond the homepage, every service page should carry a Service entity with areaServed, provider, and offers. Every service page should also carry FAQPage markup with three to five real questions answered in two to four sentences each. FAQ schema is, in practice, the highest-leverage structured data for AI citation, because LLMs naturally output question-and-answer formats. When the AI generates a recommendation, it tends to lift the answer almost verbatim from a well-written FAQ block. That is the citation, and that is your shot at being in the answer.

Citations beyond your own site

NAP consistency, name address phone, across the citation graph is non-negotiable. The directories that matter for AI search are not the directories that mattered for Google ten years ago. Foursquare is the single most important one for ChatGPT, because it feeds the data layer. Bing Places matters next. Then Yelp, BBB, Angi, and Facebook. Trade-specific directories like HomeAdvisor, Houzz, and Porch carry weight depending on your trade. Fix the data on those eight surfaces first, then worry about the long tail.

Freshness signals matter more than they used to. Industry analysis from 2026 shows Google Business Profile listings unchanged for 30 or more days lose impression share, and AI systems prefer recently verified sources. A contractor who set up their listings in 2022 and never touched them again is sending the same "stale" signal to ChatGPT, Perplexity, and Google AI Overviews simultaneously.

The four phases of the work

Picture a roofer in Bend doing $4M a year. Six trucks, mostly residential reroofs, some new construction. They want to show up when a homeowner asks ChatGPT "best roofer in Bend." This is not a checklist they finish in a weekend. It is a body of work that splits into four phases, each of which has real depth.

1. Schema architecture and trade subtype. Move off generic LocalBusiness markup to the right schema.org subtype, wire sameAs across the citation graph, attach AggregateRating mirrored to visible reviews, and align everything with the rendered HTML. Done wrong, this gets ignored or flagged by AI systems.
2. Citation graph cleanup. NAP consistency across Foursquare, Bing Places, Yelp, BBB, Angi, Facebook, and trade-specific directories. Foursquare alone is the single highest-impact source for ChatGPT, and most contractor listings on it are wrong, stale, or unverified.
3. Service page and FAQ build. A Service entity per offering, a FAQPage block per service page, and FAQ content written in the way LLMs actually quote. This is more writing work than markup work, and it is where most contractor sites fall apart.
4. Local content engine. Neighborhood-level and climate-specific content that gives AI systems something to cite when answering "best roofer in Bend" or "what roof handles Central Oregon weather." This is the long tail and the part that compounds over quarters, not weeks.

Each phase has tradeoffs we will not detail here. Phase one alone has a dozen judgment calls (which subtype if you do multiple trades, how to handle service-area versus storefront, how to avoid markup that triggers spam filters) that are not worth working out from scratch.

One contractor's actual outcome

"After we restructured the schema and cleaned up Foursquare and Bing, we started seeing the company name surface in ChatGPT recommendations within about six weeks. The hard part was being patient through the first month when nothing seemed to move."

That is not a guarantee, that is a pattern we see across contractors who do the work in the order above. The contractors who do not see results are usually the ones who did three of the six steps and stopped because it did not feel exciting.

The skeptic's question, answered

"How many homeowners are actually asking ChatGPT for a contractor in 2026?" Fair question. The honest answer is: enough that ignoring it is no longer a reasonable position, but not so many that this single channel will replace your phone. Gartner's February 2024 prediction was that traditional search volume would drop 25 percent by 2026 due to AI chatbots, and the trend line through this spring's data is on pace. AI search adoption among consumers asking for local services jumped from 6 percent to 45 percent in roughly twelve months per BrightLocal-attributed data. The window to be cited before the category fills up is closing.

What to do this week

If you do nothing else this week, do one thing: search your trade plus your city in ChatGPT, Perplexity, and Google AI Overviews. Write down which businesses get named. Cross-reference those names against your Google ranking. The gap between "I show up on Google" and "I get named by ChatGPT" is the work that needs doing. Once you can see the gap clearly, the fix is obvious. If the AI is naming three contractors in your city and yours is not one of them, that is your next quarter's project, and the cost of waiting is rising every month.

Frequently Asked Questions

How do I get my contracting business to show up on ChatGPT?

ChatGPT pulls local business data primarily from Foursquare and Bing, then runs a live search and re-ranks. To show up, you need a verified Foursquare and Bing Places listing, consistent NAP across major directories, and structured data on your site that an LLM can extract cleanly.

Is Google Business Profile enough for AI search visibility?

No. Google Business Profile drives Google AI Overviews but does not feed ChatGPT. Foursquare, Bing Places, Yelp, BBB, and Angi feed the broader AI ecosystem. A contractor who only optimizes Google Business Profile is invisible to a homeowner using ChatGPT or Perplexity.

What schema markup should a contractor website use?

Use the specific schema.org subtype for your trade: RoofingContractor, HVACBusiness, Plumber, Electrician. Add a Service entity per offering, a FAQPage block on each service page, and an AggregateRating tied to visible reviews. Format everything as JSON-LD.

How long does it take to start showing up in AI search results?

In our experience, the first signal usually appears within 30 to 60 days after schema markup, directory cleanup, and FAQ content land. Consistent visibility takes longer because LLMs sample across multiple sources before stabilizing recommendations.

How is AEO different from SEO?

SEO optimizes for blue links on a search results page. AEO optimizes for being the answer an AI generates. The mechanics overlap on technical fundamentals but diverge on what counts as a citation: SEO rewards rankings, AEO rewards being the source the AI quotes.