We took four HVAC companies currently running Google ads in Northern Virginia and ran each one through a structured AI visibility audit across ChatGPT, Claude, Perplexity, and Gemini. The results were striking — and consistent enough to say something definitive about where local service businesses stand with AI search right now.
The companies we tested
We picked companies actively competing for HVAC customers in the NoVA market: F.H. Furr (a regional chain with 45+ years in the DC metro area), McCarthy Services (advertising specifically for Alexandria HVAC), Brennan's Heating & Air (family-owned since 1979, Woodbridge-based), and Climatic Heating and Cooling (serving Leesburg and Ashburn for 20+ years).
All four are spending on Google. All four have real customer bases. Here's how they score with AI.
The results
| Company | V-Score | ChatGPT | Claude | Perplexity | Gemini | |---|---|---|---|---|---| | F.H. Furr | 48 | 43 | 53 | 43 | 53 | | McCarthy Services | 43 | 43 | 43 | 43 | 43 | | Brennan's Heating & Air | 25 | 14 | 57 | 0 | 28 | | Climatic Heating & Cooling | 25 | 43 | 28 | 0 | 28 |
The highest score is 48. The category average is 35. Every single company is rated "Low visibility."
What the models actually say
Not one of these companies appeared in AI responses to the queries that actually drive new customer acquisition. But it's worth showing what the models do say, because that's where the real story is.
"Best HVAC services near me" — here are the actual responses:
ChatGPT: "To find the best HVAC services near you, consider checking out companies like Aire Serv, One Hour Heating & Air Conditioning, and Trane. You can also use platforms like HomeAdvisor or Angie's List to read reviews and compare local providers."
Claude: "To give you specific local recommendations, I'd need your city or zip code. That said, for finding top-rated HVAC services in your area, check Angi, HomeAdvisor, or Thumbtack — they show verified reviews and let you compare quotes. Well-known national chains like One Hour Heating & Air Conditioning and Aire Serv have locations across the U.S. if you want a trusted brand."
Perplexity: "If you're in New York City or nearby, a few strong options are Premier HVAC Services for 24/7 emergency repair, 212 HVAC for residential and commercial work in NYC..."
Gemini: "For the best HVAC services near you, start by checking Google Maps or Yelp for highly-rated local companies. Look for businesses with many positive customer reviews and industry certifications like NATE."
Perplexity assumed New York City — because it doesn't know the user's location, it defaulted to its best guess. The other three either sent the user to aggregators or named national chains. Not a single NoVA company was mentioned.
The pattern for "emergency plumber in Northern Virginia" was similar. Claude recommended Roto-Rooter, Michael & Son Services, and John C. Flood. Perplexity cited specific phone numbers for John C. Flood, Beacon Home Services, and Roto-Rooter Fairfax. Gemini mentioned F.H. Furr — the only instance across all queries where one of our four companies appeared in a discovery context. It was the one query where F.H. Furr's scale gave them enough recognition to surface.
The only queries where any of the four companies appeared consistently were branded ones: "F.H. Furr HVAC reviews," "McCarthy Services customer service contact," "Brennan's Heating & Air reviews." They show up when you already know their name. For everything else, AI models point somewhere else entirely.
That's a meaningful distinction. Google ads reach people who are actively searching. If someone searches "HVAC repair Northern Virginia" and clicks an ad, they've found you through your own spend. But when that same person asks ChatGPT for a recommendation first — which is increasingly common, especially for higher-consideration purchases like HVAC replacement — these companies don't exist.
The one exception worth noting
Brennan's was the only company that appeared in a genuine category query. When Claude was asked "top rated heating and air companies in Virginia," it mentioned Brennan's Heating & Air first in its Northern Virginia recommendations.
The reason is likely their editorial footprint. Brennan's holds a Northern Virginia Magazine "Best Home Experts" designation — the kind of third-party recognition that gets into AI training data and gives a model something credible to cite. Their family-owned-since-1979 story also gives Claude a clear, specific reason to recommend them.
But that visibility only appeared on Claude. Perplexity scored them zero — a reflection of how their website performs in live retrieval. The lesson: editorial recognition helps, but it needs to be backed by strong search performance to work across all models.
Why the larger companies aren't winning
F.H. Furr and McCarthy scored higher overall, but only because brand name recognition inflates the branded query scores. Their category query performance was identical to the smaller companies: zero across the board.
This is the trap. A company that has spent decades building Google rankings and ad presence has strong brand visibility among people who already know them. AI search works differently — it's serving discovery queries, not brand queries. The signals that make you rank in Google ads (budget, landing page relevance, bid strategy) have almost no relationship to whether an AI model recommends you unprompted.
What these companies should do
The path from invisible to visible in AI search isn't mysterious, but it requires different work than Google ads.
For the larger companies (F.H. Furr, McCarthy): The gap is category association. You need to be mentioned — by name, for specific query types — in the editorial content AI models retrieve from. That means getting into "best HVAC in Northern Virginia" roundups from regional publications, not just maintaining your own website.
For the smaller companies (Brennan's, Climatic): The Perplexity zero is the most actionable signal. Perplexity uses live web retrieval, so a score of zero means your site isn't surfacing for relevant category queries in search. Fixing that — through content that targets "best HVAC in [your service area]" queries and building local editorial citations — will lift your Perplexity score and carry over to the other models.
For all four: the review volume on trade aggregators like Angi and HomeAdvisor is likely the fastest lever. These platforms rank well for local HVAC queries and get retrieved by AI models. More reviews, more consistent NAP data, more presence on those platforms = more AI visibility.
The bigger picture
If you're a local service business spending on Google ads, you're paying to be found by people who are already searching the way Google expects. A growing share of high-intent customers are starting with an AI recommendation instead. Right now, for Northern Virginia HVAC, not a single active advertiser is capturing that channel.
That's a gap — and it won't stay open for long.
Want to see how your business scores? Run a free AI visibility audit and get your V-Score across all four models.