Generative engine optimization · Free tool
How to check if ChatGPT recommends your business
Most business owners we talk to have never actually typed their own business into ChatGPT. They've googled themselves plenty, checked where they land on their Google Business Profile, maybe even run a speed test on their site. Asking the AI directly just never crossed their mind, because it feels like a strange thing to do and there's no obvious box to type it into.
We got sick of watching people guess, so we built something that asks for them.
Some people call this AEO, answer engine optimization. We call it citation engineering, because we're chasing a name in the answer rather than a spot in the rankings.
Here's how it works. You hand over your business name and city. It figures out what you actually do, then pulls five real questions, the kind a potential customer actually types when they're hunting for a business like yours nearby. Then it fires those at ChatGPT and Perplexity and checks whether your website shows up anywhere in what comes back, not just whether your name gets dropped, but whether the model actually cites your site as a source. Twenty to thirty seconds, and the results land right there on the same page.
We're not peddling some report you wait a week for. You watch the questions run live and see exactly what the model said and where it pulled that answer from.
What a "no" actually looks like
We ran this exact test on ten independent med spas in Los Angeles, real clinics, not chains, grabbed straight from a Google search. We asked ChatGPT and Perplexity to name a med spa in the city, and neither model cited a single one of the ten, not as a partial mention, not buried three lines down in a longer answer. Zero, from both models.
That's what surprises owners when they run it themselves. They figure invisibility means ranking low, showing up on page three of some AI answer nobody scrolls to. It's not that. The model either has a small set of sources it trusts for that question, or it doesn't. If your site isn't one of them, you don't rank low, you're not there at all.
We've seen this pattern hold everywhere we've run it. Day spas got named twice as often, five out of ten instead of zero, but most of those wins went to hotel-attached properties the model already trusted for other reasons rather than to independent local spas with solid reviews and a clean site. Pull those out and the independent day spas land right where the med spas did. The shape of the problem repeats: most local wellness businesses are running search-optimized websites into a system that was never reading those websites in the first place.
That's the part that's hard to sit with if you've sunk years into SEO. Meta tags, page speed, keyword-stuffed service pages, none of it's wrong exactly, it's just answering a question nobody's asking anymore. The model isn't crawling your homepage and deciding you're a good fit. It's checking whether anyone else already vouched for you somewhere it trusts.
Why we built this instead of just telling you
We could write another post about how AI visibility matters. You've probably read five of those this month already. What we haven't seen much of is a way to actually check your own situation without hiring someone or learning to run these queries by hand.
So the check does exactly what we do manually for every audit: it runs the real question and reads back exactly which sources the model cited instead of you. You get what an agency would charge diagnostic money for, and you get it before you've spent anything.
There's also a reason so few owners run this check themselves, even once they know it's possible. Typing your own business into ChatGPT and getting nothing back feels personal in a way a low SEO score never does. A number on a dashboard is abstract. Watching an AI describe your city's spa scene and skip you entirely is not. We built the check to be fast partly so that moment passes quickly, and partly so you can immediately see what to do with it instead of sitting with the bad news.
What to do once you see the answer
If the check comes back clean and your business is already cited, that's rare enough to screenshot. Most of the time it comes back the way our Los Angeles test did, and the useful next step isn't touching your website first.
Look at what the model cited instead of you. In our testing, it's almost never a competitor's homepage. It's a directory, a review aggregator, a local roundup, the kind of page you'd find if you searched "best [your category] in [your city]" yourself. We wrote up the specific directories that keep showing up across wellness categories, and it's a short, repeatable list, not a hundred platforms you'd need to chase.
Getting listed there, with your name, address, and phone number matching exactly across every one, does more for your citation odds than another round of schema markup or a redesigned homepage. We tested that claim directly too, and the results didn't favor the on-page fixes agencies usually sell first.
None of this requires guessing. Run the check, read what the model actually cited, and go fix the gap between those sources and your own listings. That's the whole method, and it's the same one we use before we touch anything else on a client's site.