H1: How Does ChatGPT Pick Local Businesses for Trades?
When a homeowner asks ChatGPT for a plumber, the system does not search the way Google does. It does not return ten blue links ranked by keywords and backlinks. It runs a multi-step pipeline that cross-references several data sources, verifies the business exists, checks what other people say about it, and then commits to two or three recommendations.
SOCi's 2026 Local Visibility Index analyzed over 350,000 business locations across 2,751 brands. ChatGPT recommends 1.2% of local businesses. Gemini recommends 11%. Perplexity recommends 7.4%. The businesses that make the cut share a specific set of signals. The businesses that do not are missing one or more of them.
For trades businesses, the stakes are high. BrightLocal's 2026 survey found that 45% of consumers now use AI tools to find local services, up from 6% one year earlier. A growing share of your potential customers are asking AI for a recommendation. Almost none of them are seeing your name.
H2: The Five-Step Recommendation Pipeline
ChatGPT's local recommendation process is closer to a verification workflow than a search query. Here is what happens when someone types "best HVAC contractor near me."
H3: Step 1. Query Expansion
ChatGPT takes the original prompt and fans it out into multiple sub-queries. "Best HVAC contractor near me" becomes several parallel lookups: "HVAC contractor in [city]," "air conditioning repair [city]," "furnace installation [city] reviews," "licensed HVAC contractor [city]." This process is called query fan-out, and it is how most major AI assistants now handle local questions.
Each sub-query targets a different data source. Some hit Google Business Profile data. Some hit web search results. Some hit review platforms. The AI gathers information from all of them in parallel, then stitches the results together.
For trades businesses, this means your category selection on Google Business Profile matters more than ever. A contractor listed as "HVAC Contractor" gets pulled into HVAC sub-queries. The same contractor listed as "Construction Company" does not, even if they offer the same services.
H3: Step 2. Entity Resolution
After gathering results from multiple sub-queries, ChatGPT needs to figure out which businesses are real. It cross-references business names, addresses, and phone numbers across the sources it pulled. If "Joe's Plumbing" on Google Business Profile, "Joe's Plumbing" on Yelp, and "Joseph's Plumbing LLC" on Angi all share the same address and phone number, the AI may merge them into a single entity. Or it may treat them as three separate businesses.
This is where NAP consistency becomes critical. The more platforms your business information matches on, the more confidently the AI can resolve your entity. Inconsistencies lower the confidence score. Low confidence means the AI skips your business rather than risk recommending a company it cannot verify.
Research from SOCi shows that AI systems cross-reference at least 10 platforms for each local business. Google Business Profile, Yelp, Facebook, Bing Places, Apple Maps, Yellow Pages, BBB, Foursquare, Nextdoor, and Angi are the core set. Trades businesses listed on all 10 with identical NAP data get resolved as a single, high-confidence entity. Businesses listed on five with minor variations get fragmented across the AI's understanding.
H3: Step 3. Trust and Reputation Scoring
Once the AI has resolved your business as an entity, it evaluates trust and reputation. This is where most trades businesses get eliminated.
ChatGPT-recommended businesses average 4.3 stars, according to SOCi's 2026 data. Businesses near 3.4 stars with response rates below 5% do not get ranked lower. They get excluded from the candidate pool entirely. The AI treats low ratings combined with low engagement as a signal that the business may be inactive or unreliable.
AI tools analyze the text inside reviews, not just the star count. They look for service-specific language. "Fixed my burst pipe fast" tells the AI this is a plumber who handles emergencies. "Showed up on time and explained everything" signals professionalism. A review that says "great company" tells the AI nothing useful and gets weighted lower.
Response rate matters independently of rating. A business with 60 reviews and 60 owner responses signals an active, engaged operation. A business with 200 reviews and 5 responses signals a company that showed up, collected stars, and stopped paying attention. AI systems treat the first profile as more trustworthy.
H3: Step 4. Content and Schema Evaluation
After trust scoring, the AI evaluates what your website actually says about your business. This is where schema markup and content structure come in.
Sites with proper LocalBusiness schema get cited 3.2 times more often in AI responses, according to 2026 research. Schema gives the AI a machine-readable version of your business information. Business name, address, phone, hours, service area, aggregate rating, and review count all get parsed without the AI having to interpret your page layout.
The AI also evaluates content structure. Pages that open with a direct answer to a question get extracted more reliably. Pages that bury the answer under three paragraphs of company history get skipped. FAQ sections give the AI answer-shaped blocks it can pull directly into its response.
For trades businesses, this means every service page should answer the question a customer would ask. "How much does furnace replacement cost in [county]?" with a price range in the first sentence. "Do you offer emergency plumbing?" with a yes or no in the first sentence. AI extracts from the top of the section, not the bottom.
H3: Step 5. Synthesis and Recommendation
In the final step, the AI combines all the signals into a ranked list and generates a natural-language recommendation. It picks two or three businesses, describes what makes each one worth considering, and cites the sources it used.
The synthesis step is where the AI's personality shows. ChatGPT tends to favor businesses with strong web presence and publisher mentions. Gemini relies more heavily on Google Business Profile data and local signals. Perplexity favors niche and regional sites with detailed content. Each AI engine has a slightly different bias, which is why the same business might appear in Gemini's answer but not ChatGPT's.
This also explains the volatility. Research from F9XR found roughly 85% volatility in AI recommendations, meaning the businesses AI recommends can shift significantly between queries. A business that shows up in today's answer might not appear tomorrow. Ongoing signal maintenance matters more than a one-time optimization.
H2: What This Means for Trades Businesses
The businesses that win AI recommendations are not the ones with the biggest ad budgets or the most backlinks. They are the ones with the cleanest entity signals. Complete Google Business Profile, consistent NAP across 10+ directories, reviews with service-specific text and 100% response rate, LocalBusiness schema on the website, and content structured for extraction.
Most trades businesses have none of these dialed in. That is the opportunity. A plumber who fixes all five signal categories this month has a realistic shot at becoming one of the 1.2% before the competition catches up.
The cost of inaction compounds. Every month that consumers shift more of their local search to AI tools, the businesses absent from those answers lose market share they may not recover. BrightLocal's data shows a 39-point swing in consumer AI adoption in a single year. The next year will move faster.
H2: How to Get Picked. A Trades Business Action List
Complete your Google Business Profile
Fill every field, upload at least 20 photos, write service descriptions, and turn on all applicable attributes. Update the profile at least monthly to keep it active. This is the single most important signal for every AI system.
Standardize your NAP across directories
Pick one name, one address, one phone. Push that combination to Google, Yelp, Facebook, Bing Places, Apple Maps, Yellow Pages, BBB, Foursquare, Nextdoor, and Angi. Do not let abbreviations or formatting differences fragment your entity.
Build review velocity with service-specific text
Ask every satisfied customer for a review. Ask them to mention the specific service performed. "Fixed my leaking water heater" is worth more than "great service." Respond to every review within 48 hours, including the bad ones.
Add LocalBusiness schema to your website
WordPress users can add it through Rank Math or Yoast. For custom sites, add JSON-LD to the homepage and every service page. Include business name, address, phone, hours, service area, and aggregate rating.
Write answer-shaped content on every service page
Open with a direct answer to the question a customer would ask, put the price range in the first sentence, include the service area in the first paragraph, and add an FAQ block at the bottom of the page. AI extracts from the top of the section, not the bottom.
FirstCall's free AI visibility audit checks all of these signals for your business and flags which ones need attention.
FAQ
How does ChatGPT pick local businesses?
ChatGPT runs a five-step pipeline: query expansion (fanning the prompt into sub-queries), entity resolution (cross-referencing NAP across 10+ platforms), trust and reputation scoring (reviews, star ratings, response rates), content and schema evaluation (checking your website for machine-readable information), and synthesis (combining all signals into a ranked recommendation).
What percentage of local businesses does ChatGPT recommend?
ChatGPT recommends 1.2% of local businesses, according to SOCi's 2026 Local Visibility Index, which analyzed over 350,000 business locations. Gemini recommends 11%, and Perplexity recommends 7.4%.
Do Google rankings help with ChatGPT recommendations?
Google rankings help but do not guarantee AI visibility. SOCi found only a 45% overlap between businesses that rank well in Google local search and businesses that appear in AI recommendations. More than half of Google map pack winners are absent from AI answers.
What star rating does my business need for AI recommendations?
ChatGPT-recommended businesses average 4.3 stars. Businesses near 3.4 stars with response rates below 5% get excluded entirely. Responding to every review matters as much as the star count itself.
How often do AI recommendations change?
AI recommendations have roughly 85% volatility, according to F9XR research. The businesses AI recommends can shift significantly between queries. Ongoing signal maintenance, monthly visibility checks, and consistent review generation matter more than a one-time optimization.
Run your free AI visibility audit at firstcallvisibility.com.
<!-- FAQPage JSON-LD -->