HVAC Website Trust and AI Search: Lessons with Shaurya Jain

From A Chat with Chat, episode 3, recorded October 9, 2026. Watch the full interview and read the transcript.

A website can attract attention and still leave a homeowner unsure whom to trust or how to get help. In our first guest interview, Aaron Husak and ChatGPT spoke with Shaurya Jain, founder of Honest HVAC Marketing Agency, about that gap.

Shaurya brought research on HVAC websites and customer reviews, along with an account of receiving inquiries through ChatGPT. The strongest practical thread was how contractors present their business, help a visitor contact them, and verify whether marketing is producing real inquiries.

A website-feature audit is a starting point

Shaurya described using Python and Claude to examine roughly 1,300 HVAC websites. His checks covered contact options, trust signals, measurement, and usability on a phone. He explained that some candidate sites could not be examined because they blocked the automated access, so they were excluded.

Among the gaps he discussed were missing alternatives to calling, limited presentation of customer reviews, and a lack of identifiable crew or work photos. These are his reported research findings. The interview did not independently validate every automated assessment.

More importantly, the study examined website features. It did not establish how many leads those sites generated or measure a conversion improvement after a repair. A missing feature is worth investigating. Its effect on bookings needs a separate test.

Give homeowners a reason to trust the people arriving

Aaron connected the discussion to the realities of home service: a customer is choosing people who will enter their home. Real team photos, equipment or project examples, and relevant customer feedback can help someone evaluate the company behind the page.

Review your own site from that customer's perspective. Can someone understand who performs the work? Are qualifications and service coverage accurate? Do project examples explain the problem and the work actually performed? Use photos you have permission to publish and protect customer privacy.

Original evidence gives readers something concrete to consider. It should not be presented as a guaranteed mechanism for winning an AI recommendation.

Make contact options work in practice

The interview's website discussion offers a useful inspection checklist: a visible phone number, a working mobile call link, a clear form or booking path when appropriate, and readable information on smaller screens.

Check delivery as well as appearance. A form on a page is not proof that messages reach the team. A booking button is not proof that a usable appointment can be selected. Verify the complete path, then confirm who responds to the inquiry.

Only add chat, texting, or another contact method if someone owns it. Creating more ways to reach the company can create more unanswered requests when the handoff is unclear.

Treat a ChatGPT inquiry as something to investigate

Shaurya described a one-page website receiving inquiries through ChatGPT despite little intentional SEO work. That is his account of what happened, not a demonstrated recipe for generating leads without SEO.

The useful follow-up is attribution: did the person report finding the company through ChatGPT, and can any identifiable referral be connected to an inquiry? Those are different types of evidence. Capture the question and answer when possible, and keep self-reported discovery separate from tracked visits.

Our discussion also explored location and repeated searches. A consistent answer in one session does not establish what every person in the named city will see. Document test conditions before comparing results.

Respond to reviews as an operator

The conversation turned to negative reviews and the experience behind them. A public response should show that the company listens and offers a constructive way to resolve a problem. Avoid exposing customer details or turning the reply into an argument.

Look for recurring concerns that deserve operational attention: expectations, explanations, communication, or follow-through. A response cannot substitute for fixing the underlying issue.

One useful action after the interview

Pick one customer path and inspect it from discovery through response. Identify one specific gap, correct it, and compare actual inquiries before and after while noting other changes. Keep qualified inquiries and jobs booked separate from page visits or AI mentions.

Explore the show and available episodes, read our Gemini test recap, or talk with Sequoia GEO about your website and inquiry process.