What AI engines tell people who ask “How do I find reliable pool contractors in CT?”

AI Visibility for CT Pool Contractors

This issue examines how the major AI engines answer a question that Connecticut homeowners ask before they hire: “How do I find reliable pool contractors in CT?” I ran that exact prompt through six AI engines and recorded which Connecticut pool companies each one named, which it skipped, and where its details held up or fell short. Below, I lay out which companies the engines name, the patterns that repeat across all six, and what a Connecticut pool business can do to be named more often.

Welcome to the CT Brief

This is the first issue of the CT Brief, a monthly look at AI Visibility in Connecticut. Each issue takes one question that buyers in a specific Connecticut industry ask, runs it through the six AI engines a buyer is likely to use (ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity, and Claude), and reports what those engines say, and do not say, about Connecticut businesses. The premise is concrete. More buyer research now happens inside AI engines than on a page of search results, and when someone asks an engine for help choosing a company, it answers with a short, curated list. A business that the engines do not name is invisible to that buyer, regardless of how it ranks on Google. Each issue tracks who gets named, who does not, and the patterns that emerge, and it closes with a specific takeaway that a business in that industry can act on. You can subscribe at the end of this piece to receive future issues by email.

The question that started this issue

For Issue 1, I picked a question that matters to every Connecticut pool contractor, builder, and installer: “How do I find reliable pool contractors in CT?”

This is the question a homeowner asks before they call anyone. It is the first practical step in a major home purchase, in plain language. And the AI engine’s answer determines which businesses get a call. For pool contractors, being included in the engine’s answer is the difference between being considered and not being considered. So: who’s being included?

I asked the same question across six AI engines:

• ChatGPT (manual run in a fresh session)
• Google AI Overview (incognito Google search)
• Google AI Mode (Google search → AI Mode)
• Gemini (manual run in a fresh session)
• Perplexity (manual run in a fresh session, incognito)
• Claude (manual run in a fresh chat session)

For each engine, I captured the full response, the businesses named in the answer, and the sources the engine cited. The questions I went in with: which Connecticut pool contractors appear across multiple engines? Which appear in only one? What sources is each engine leaning on? And what does the variation tell a contractor about how to be visible?

One caveat on reading the results: this is a snapshot. AI engines do not return identical answers every time, and what they say shifts as the underlying web changes. The same question asked next month may return a somewhat different set of businesses. Treat the specific names and counts here as directional. They are a picture of how the engines were answering when I ran it, not a permanent ranking.

Same question, six engines, six distinct answers. Here is what each one named when asked who Connecticut homeowners should call:

• ChatGPT named six Connecticut pool contractors: Pools By Murphy, EZ Aqua Pool & Patio, Juliano’s Pools, RI Pools, Pinto Pools, Rizzo Pools.
• Google AI Mode named four: Drakeley Pool Company, Lang Pools, Prospect Pools, Anthony & Sylvan Pools.
• Gemini named five: Total Pool + Patio, Anthony & Sylvan, Fiore Custom Pools, Aquatic Pool & Spa, Lang Pools.
• Perplexity named eight: Pool Doctor, Drakeley Pool Company, Poolscape Pools & Spas, Pools by Murphy, Lang Pools, Fiore Custom Pools, Total Pool + Patio, plus one apparent garbled name (more on that below).
• Google AI Overview named three: Lang Pools Inc., Pools by Murphy, Anthony & Sylvan Pools. It was the only engine to include a direct phone number for one of them.
• Claude named zero Connecticut pool contractors. Instead of recommending businesses, Claude pointed to the Connecticut eLicense Portal, the Department of Consumer Protection’s Home Improvement Guaranty Fund, the Northeast Spa & Pool Association (NESPA) member directory, and local building departments. That places the burden of selection on the buyer rather than offering picks.

That last one matters for any contractor thinking about AI visibility strategy: not every engine answers the question by recommending companies. One major engine declines to recommend. Your AI Visibility strategy needs to account for engines that operate as referees, in addition to engines that operate as recommenders.

When the same business appears in multiple engines’ answers, that’s the signal worth paying attention to. Out of 16-plus unique Connecticut pool businesses named across the five recommending engines, only six appeared in more than one engine. The rest were singletons, cited by exactly one engine and invisible in the others.

The cross-engine leaders:

• Lang Pools (Norwalk). Appeared in four of the five engines that named businesses. AI Mode, Gemini, Perplexity, and AI Overview all cited Lang Pools. Only ChatGPT didn’t. No other Connecticut pool contractor was cited as consistently.
• Anthony & Sylvan Pools (Westport). Appeared in three of five engines (AI Mode, Gemini, AI Overview).
• Pools by Murphy (Wilton/Canton). Appeared in three of five engines (ChatGPT, Perplexity, AI Overview).

Below those three, three businesses appeared in two of five engines: Drakeley Pool Company, Total Pool + Patio, and Fiore Custom Pools. Everything else was a singleton.

The takeaway from the leaderboard isn’t that Lang Pools is the “best” Connecticut pool contractor. AI engines don’t rank by quality; they synthesize from public signals. The leaderboard tells you which Connecticut pool businesses are producing the trust signals AI engines weigh when answering this question.

Here’s the most important pattern in the data, and the one that should change how every Connecticut contractor thinks about AI Visibility. The six engines cited 16-plus unique Connecticut pool businesses across their answers to the same question. Only one appeared in four of five. Only two others appeared in three of five. Ten or more were cited by exactly one engine.

What this means in practice: a Connecticut homeowner asking ChatGPT will see a substantively different recommendation set than a homeowner asking AI Mode, Gemini, Perplexity, or AI Overview. ChatGPT named six pool contractors. Of those six, only one (Pools by Murphy) appeared in any other engine. The other five ChatGPT-cited businesses appeared in none of the other engines. Those were EZ Aqua, Juliano’s, RI Pools, Pinto Pools, and Rizzo Pools. If your AI Visibility strategy is “we get cited on ChatGPT,” you are invisible to buyers using anything else.

The reverse is also true. The single most consistently-cited Connecticut pool contractor in this research (Lang Pools) appears in four engines but is invisible on ChatGPT. A buyer using ChatGPT would never hear about Lang Pools through that channel. Each engine has a substantively distinct recommendation universe. There is no shared “top Connecticut pool contractors” list.

The reason the engines disagree about who to recommend is that each one is pulling from a distinct set of trust signals. The sources cited tell the story:

• ChatGPT pulled heavily from Reddit threads, BuildZoom, BestProsInTown, HomeAdvisor, Angi, Birdeye, AllBiz. Forum discussions and third-party directory aggregators.
• AI Mode pulled from Google Business Profile data, the CT.gov licensing portal, the eLicense.ct.gov lookup tool, Houzz, and individual company websites. Google’s own structured data plus state government plus design platforms.
• AI Overview pulled from the same Google Business Profile layer plus Houzz, plus the websites of the recommended companies themselves.
• Gemini pulled from industry association sites (NESPA), the CT.gov licensing portal, an insurance industry blog, and the individual websites of the businesses it recommended.
• Perplexity pulled from a mix of all of the above and organized recommendations by Connecticut region (Wolcott, Fairfield County, central and northwest CT). The only engine to organize by geography.
• Claude pulled exclusively from regulatory and licensing infrastructure and declined to recommend specific businesses. Its sources were the Connecticut DCP, the eLicense Portal, the Bridgeport building department, and NESPA.

Different engines, different trust hierarchies. ChatGPT trusts what real people say in public forums. AI Mode and AI Overview trust what Google’s structured data layer shows. Gemini trusts what professional associations and government sources show. Perplexity blends and organizes geographically. Claude trusts regulators and tells you to do the picking yourself. That’s the mechanical reason for the fragmentation. The strategic implication is clearer once you see it: to appear consistently across engines, you need to be visible across multiple trust signal types simultaneously.

This is the part that’s worth doing something about. If you do not openly display your licenses, credentials, reviews, and other trust signals in places where AI engines look, and across enough of those places that multiple engines find you, you will be invisible to buyers using AI search. An AI Visibility Audit shows which of those places are actually working for you and which are not.

Visibility in one channel is not enough. AI engines synthesize across multiple trust signal types, and each weighs them differently. A pool contractor with a strong Google Business Profile but no NESPA membership might appear in AI Mode and AI Overview but not in Gemini. A contractor with a great industry association presence but a thin GBP might appear in Gemini but not in AI Overview. A contractor relying primarily on word-of-mouth and Reddit reputation might appear in ChatGPT but in nothing else. Measuring that across all six engines, and against up to nine competitors, is what AI Visibility Intelligence does.

The Connecticut pool contractors that appeared most often in this research share visible presence across multiple trust signal types. Those are Lang Pools, Anthony & Sylvan, and Pools by Murphy. Verifiable state licensing on the eLicense Portal. Active Google Business Profile with reviews. NESPA membership. Their own websites with clear contact information, named staff, and project portfolios. Third-party visibility on Houzz, BBB, Angi, BuildZoom. The trust signals compound across channels, and the AI engines pick that up. Concrete actions a Connecticut pool contractor (or any Connecticut home services business) should consider:

The work isn’t glamorous. None of these are new SEO concepts. What’s new is that AI engines now weight these signals as heavily as Google traditionally has, and they each weight them differently. The pool contractor who treated GBP as a “set it and forget it” task five years ago is now disadvantaged compared to the contractor who keeps it active and complete.

One observation about Claude is worth a separate mention. Claude was run from a signed-in account, and its response included an aside referencing one of Telstar’s pool industry clients, pulled from account memory unprompted. The other engines didn’t do anything comparable.

This is account-based personalization, distinct from the geolocation-based personalization that Perplexity and AI Overview both displayed. The substantive content of Claude’s response, its emphasis on regulatory verification and lookup tools rather than direct business recommendations, would likely be similar for any user. But the specific aside reflected the user account context.

For Connecticut contractors, the implication is small but worth knowing: when a customer uses Claude to research your category, they may get a response shaped by their own account history. This is a new variable in the AI Visibility picture and one that may grow more significant as AI engines lean more heavily on personalization.

Issue 2 will use the same single-prompt, deep-analysis format for a different Connecticut industry. The exact prompt will be picked the same way, a question that’s first-step-critical for the industry and that real Connecticut buyers ask AI engines. If there’s a Connecticut industry you’d like covered in a future issue, let me know.