Generative engine optimization · Original data
Why is my business not on ChatGPT or Perplexity?
Most owners figure that if ChatGPT has heard of their business, Perplexity must have too. Same internet, same reviews, same Google listing — why would two AI models serving up the same local market end up in different places?
We put that assumption to the test. Fifty actual, independently-verified businesses across five categories — day spas, med spas, boutique salons, yoga studios, wellness retreats — in cities stretching from San Diego to Austin to Miami. Same question into GPT-4o and Perplexity's Sonar, same week, worded like a real customer would actually type it: "best [category] near me."
GPT-4o named thirteen businesses. Perplexity named nine. Seventeen got named by at least one of them. And only five — five out of fifty — showed up in both.
That's not two models zeroing in on the same local landscape. That's two models that mostly can't agree on who's even in it.
The category breakdown makes it worse
Wellness retreats had the strongest overlap in our entire test — 3 of 10 businesses named by both engines, out of 6 that got any mention at all. That's the best-case scenario, and it still means half the businesses either engine named, the other one completely missed.
Day spas and boutique salons had zero overlap. GPT-4o named three day spas, Perplexity named two, and not a single one matched. Same story in salons: three named by GPT-4o, none of them surfacing in Perplexity's answer.
Med spas didn't have an overlap problem. They had a visibility problem — zero citations from either model, across all ten businesses we tested. We covered that category collapse separately because it's its own thing. But even in categories where AI does mention real businesses, the two engines are drawing from different information nearly every time.
Why this isn't a fluke
The easy explanation is that GPT-4o and Perplexity pull from different sources when they build answers — Perplexity leans more heavily on live web retrieval, GPT-4o draws more from what it already knows plus whatever it fetches. Different retrieval, different citations. That part isn't shocking.
What's worth actually sitting with is what this means for an owner. There's no single "AI version" of your local market that you can check once and trust. If your business comes up when someone asks ChatGPT, that tells you nothing about what Perplexity tells the next person asking the same question in the same city. You could be visible to half your potential customers and invisible to the other half, and you'd have no clue which half without checking both directly.
What actually moved the needle for the businesses that got picked up by both
The five businesses that made it into both engines' answers weren't the biggest on social media or the newest kids on the block. They shared two things: they'd been around long enough to build real review histories across multiple platforms, and their name, address, and category matched everywhere someone might look them up — Google Business Profile, Yelp, their own site, local directories.
That consistency matters more than it sounds. When a model decides whether to name a business, it's essentially cross-checking whether the same facts appear in more than one place. A business with identical name and address on five listings reads as one real, stable place. A business with "Sunrise Wellness Spa" on Google and "Sunrise Spa & Wellness" on Yelp looks, to a model doing quick pattern-matching, like it might be two different, less-established spots — or it just gets skipped because the signal isn't clean enough to commit to.
What to actually do with this
Checking your visibility on one engine and calling it finished is checking half the picture, maybe less. If you've only ever typed your business into ChatGPT and seen your name appear, that's genuinely good news — but it tells you nothing about what a customer asking Perplexity, or Google's AI Overviews, or Claude is getting instead.
The fix isn't picking a favorite engine to optimize for. It's making the underlying facts about your business — name, address, category, reviews — identical everywhere a model might look, so that whichever one someone happens to ask, the same clean signal is sitting there waiting.
You can check where you actually stand across ChatGPT, Perplexity, and Claude in about thirty seconds — no signup, nothing to install. Given that even the best-performing category in our test only had a 30% agreement rate between engines, it's worth finding the gap yourself before a customer finds it for you.