How to Tell If AI Is Recommending Your Competitors Instead of You

Most owners have no idea what AI assistants say about their category, because the answer only appears when a customer asks the question — and customers don’t report back. If a competitor’s name is the one coming out of ChatGPT or Perplexity right now, you could be losing leads you’ll never see in an analytics dashboard, with no obvious signal that it’s happening. The good news is that checking this takes about twenty minutes, and closing the gap is a defined project, not a mystery.

Why you can’t tell from your normal metrics

Website analytics show you traffic that already arrived. Google Search Console shows you ranking and impression data for traditional search. Neither tool shows you what happens inside a ChatGPT conversation or a Perplexity answer, because that interaction never touches your site unless the AI includes a link and the user clicks it — and increasingly, it doesn’t include one at all.

That means a real, measurable shift in buyer behavior can be happening — prospects getting a recommendation, forming an opinion, and shortlisting two vendors before ever visiting a website — and it will be completely invisible in the tools you check every week. You have to go look for it directly.

The direct check: ask the questions your buyers ask

Open ChatGPT, Perplexity, and Google (to see AI Overviews) and ask the same questions a prospect would type when they’re trying to find a vendor in your category. Not your brand name — the generic questions. For a multi-location HVAC company, that’s things like:

• “Who are reliable commercial HVAC contractors that service multiple locations in [region]?”
• “What should I look for in a commercial roofing contractor for a multi-site property?”
• “Best equipment dealers for [category] with strong parts and service support?”

Ask each question in a fresh conversation, without prior context that would bias the answer toward you. Run each one a few times, since answers can vary between sessions. Do this across all three platforms — they pull from different sources and often name different businesses.

What to look for in the answers

Read each response for three things:

Is anyone named at all? Some answers stay generic — “look for a licensed, insured contractor with good reviews” — without naming specific businesses. If that’s what you’re getting across the board, the opportunity is wide open in your category and worth acting on before a competitor claims it.

Who gets named, and how often? If the same one or two competitors show up across multiple questions and multiple platforms, that’s not a coincidence — it means their content, reviews, and structured data are consistently legible to these systems in a way yours isn’t yet.

What does the AI say about the named businesses? Look at the specific claims. Often it’s things like service area, certifications, years in business, or a specific specialty. That tells you exactly what kind of information these systems are finding and trusting — which is a direct guide to what your own pages are missing.

Digging one layer deeper

If a competitor is getting named, ask a natural follow-up in the same conversation: “why did you recommend them” or “what makes them a good fit for multi-location properties.” The AI will often explain its reasoning in terms of what it found — clear service descriptions, review volume, a specific certification. That explanation is close to a checklist of what your own site is currently missing.

It’s also worth searching your own business name directly. If the AI gives a vague or thin answer about you — or gets something wrong, like an outdated service list or the wrong locations — that’s a legibility problem on your end, not a reputation problem. These systems are reading what’s actually published about you, and if it’s outdated or sparse, that’s what comes back.

Closing the gap

Once you know who’s getting named and why, the fix is concrete, not abstract:

1. Match the specificity of the businesses that are winning. If a competitor’s edge is a clearly stated service list and coverage area, your pages need the same level of plain, factual detail — not more marketing language.
2. Fill in structured data. LocalBusiness, Service, and FAQ schema on your key pages gives these systems a clean, machine-readable version of the facts, rather than making them infer everything from prose.
3. Address whatever the AI cited as the reason. If reviews were the deciding factor, a stale review profile is the first thing to fix. If certifications were mentioned, make sure yours are stated clearly and prominently.
4. Re-run the same questions in a few weeks. This isn’t a one-time audit. AI systems refresh what they’ve indexed on their own schedule, so treat this check as a recurring part of how you monitor your market position, not a single pass.

Checking this regularly and building the content and schema to close the gap is a defined piece of work, and it’s part of what we set up for clients alongside their paid search and paid social programs. If you want a look at what AI is currently saying about your category, book a call.