We Tracked Where AI Chatbots Send Homeowners for Plumbing and HVAC Recommendations. The Results Were Wild.
I hope you enjoy reading this blog post. If you want us to just build your AI agents for you, click here.
We asked ChatGPT, Gemini, and Perplexity one simple question across 10 U.S. markets:
“Who’s the best plumber near me?”
Then we did it again for HVAC. Then electrical. Then roofing.
What came back was fascinating, inconsistent, and a little terrifying if you run a home services company.
Some brands showed up everywhere. Some showed up nowhere.
The reasons had almost nothing to do with traditional SEO.
Welcome to the new battleground for contractor visibility: AI citation tracking home services. It’s the natural extension of everything we cover in our broader guide to getting found by AI search.
Key Takeaways
- AI chatbots are already recommending contractors by name in response to homeowner queries, with no ads or paid placement involved.
- Google’s top organic results overlapped with AI recommendations less than 35% of the time.
- The companies getting cited share specific traits: broad directory presence, consistent NAP data, and reviews across multiple platforms.
- In 3 of 10 markets, Perplexity recommended a company with no website at all, pulling entirely from Yelp and BBB data.
- Contractors who build the right digital footprint now will own the next wave of inbound leads.
The Experiment
We selected 10 mid-size U.S. metros: Charlotte, Tampa, Denver, Phoenix, Columbus, and five others.
Markets where home services companies compete hard.
We prompted each AI assistant with variations of real homeowner queries:
- “Best plumber in [city]”
- “Who should I call for AC repair in [city]?”
- “Reliable HVAC company near [zip code]”
- “Emergency plumber [city] recommendations”
We logged every company named, every source cited, every link provided.
Then we cross-referenced against Google’s top 10 organic results for the same queries.
Methodology Notes
All prompts were run within the same 72-hour window in fresh, logged-out sessions.
Location services were disabled; we specified cities and zip codes directly in each prompt.
Each query was run three times per platform to check for consistency.
Results were logged in a shared spreadsheet with timestamps.
Google vs. AI Recommendations: A Side-by-Side Look
The overlap between Google’s #1 result and AI recommendations? Less than 35%.
Here’s what that looked like in practice:
| Market | Google #1 Result | ChatGPT Pick | Gemini Pick | Perplexity Pick |
|---|---|---|---|---|
| Phoenix | Southwest Plumbing Co. | Canyon State Plumbing | Parker & Sons | Canyon State Plumbing |
| Tampa | Roto-Rooter | Climate Design | Roto-Rooter | Ben Franklin Plumbing |
| Denver | Blue Sky Plumbing | Brothers Plumbing | Blue Sky Plumbing | Baker Home Energy |
| Columbus | Apex Plumbing | Eco Plumbers | Eco Plumbers | Apex Plumbing |
In Phoenix, the #1 Google result wasn’t mentioned by ChatGPT or Perplexity once.
Canyon State Plumbing, which sits on page 3 of Google, was the first recommendation from two out of three AI platforms.
How Each AI Platform Differed
This is where it gets interesting.
Gemini leaned heavily on Google Business Profile data and Google Reviews. If you dominated Google’s local pack, Gemini usually noticed you.
ChatGPT pulled more from aggregated directory data and community mentions. It cited sources about 40% of the time, often linking to Yelp or Angi profiles.
Perplexity was the most source-transparent, citing specific URLs with nearly every recommendation. It also surfaced companies from BBB and Thumbtack that the other two ignored entirely.
The takeaway? Each platform synthesizes trust differently. Winning on one doesn’t guarantee visibility on the others.
AI Citation Tracking Home Services: What Gets You Cited vs. What Gets You Skipped
Patterns emerged fast.
Companies That Got Cited
Across the companies AI recommended, the median profile looked like this:
- Present on 17+ directories with identical business info
- 230+ Google reviews with a 4.5+ average
- Mentioned by name on Reddit, Nextdoor, or local Facebook groups
- Had LocalBusiness schema markup and FAQ structured data on their sites
- Published location-specific service pages with real neighborhood-level content
Companies That Got Skipped
- Thin websites with no location pages
- Inconsistent names across directories (“Smith Plumbing” vs. “Smith Plumbing LLC” vs. “Smith’s Plumbing Co.”)
- Few or no reviews outside Google
- Zero mentions on community platforms or forums
AI assistants are synthesizing trust signals from across the entire internet. Not just your website.
Your homepage could be beautiful. Your Google Ads could be crushing it.
But if the rest of your digital footprint is thin or inconsistent? AI skips right over you.
Inconsistent listings also route real calls to disconnected numbers or old locations.
→ Curious how many calls you’re already losing? Run your free Phone Leak Analysis and find out in 60 seconds.
How to Build the Digital Footprint AI Actually Pulls From
Here’s your tactical playbook.
1. Audit Every Directory Listing
Start with the big ones: Google Business Profile, Yelp, BBB, Angi, HomeAdvisor, Thumbtack.
Then hit the second tier: Nextdoor, Porch, Houzz, local chamber directories.
Make your business name, address, phone number, and service descriptions identical everywhere.
2. Build Location Pages That Teach
AI assistants love pages that answer specific questions about specific places.
Instead of one generic “Service Areas” page, create individual pages for each market you serve.
Include real details: neighborhoods, common issues in that area, pricing context.
We broke this approach down in depth in our location page SEO guide.
3. Get Mentioned in Community Conversations
Reddit threads, Nextdoor recommendations, and Facebook group discussions are gold.
AI models weight real human recommendations heavily.
Encourage your happiest customers to mention you by name when neighbors ask for referrals online.
Free. And it compounds.
4. Stack Reviews Across Multiple Platforms
Google reviews matter. But so do Yelp, BBB, and Facebook reviews.
AI assistants pull from all of them.
Aim for review depth AND breadth. 500 Google reviews with zero Yelp presence looks lopsided to an AI model.
5. Monitor Your AI Visibility Monthly
According to Backlinko, 78% of businesses audit their SEO once a year or less.
That mindset will be fatal in the AI citation era.
Prompt ChatGPT, Gemini, and Perplexity for recommendations in your market every month.
Track who shows up. Track when you show up. Track which sources get cited.
Then adjust your footprint accordingly.
The Final Word
Here’s where this is heading.
A homeowner’s water heater fails at 9 PM.
They grab their phone. Instead of Googling, they ask their AI assistant.
The assistant recommends three companies. Yours is one of them.
Or it’s not. This is exactly why a dedicated plumbing answering service or an always-on HVAC call line matters more than ever, the AI is sending you the lead, but only if someone actually answers.
That moment is already happening thousands of times a day.
The gap between companies that show up and companies that vanish is only going to widen.
The playbook isn’t complicated. Consistent listings. Rich reviews across platforms. Community mentions. Real location content.
But it requires intentionality.
And it requires someone answering that 9 PM call when the AI sends it your way.
Call Jill, our voice AI agent. She’ll walk you through how AI-driven leads connect to your phones and get you set up in minutes.
Your future customers are already asking AI who to call.
Make sure the answer is you.