Generative engine optimization · Original data
We asked ChatGPT and Perplexity for a boutique salon in Miami. Perplexity named none.
Not one out of ten. We ran the same query, "what are the best boutique hair salons in Miami, list 10," through GPT-4o and Perplexity Sonar as part of a wider 50-business AI visibility check across wellness categories. GPT-4o came back with three real names: Avant-Garde Salon And Spa, Oren Salon, Danny Jelaca Salon. Perplexity's list didn't overlap with GPT's at all, and it didn't include any of the ten salons we were checking for, meaning it recommended zero of the actual boutique salons operating in Miami.
That's not some minor model quirk. It's a complete blank on an entire local category, and it's worth really sitting with, because most owners hear "AI doesn't know me" and picture all the engines being equally lost. They're not. One can see you and the other can't, and you won't discover that unless you test both.
Ten businesses and one query each is a small sample, not a scientific study, so treat the exact split as directional rather than gospel about how these models behave everywhere. What holds up no matter the sample size is the pattern: GPT-4o and Perplexity built almost entirely different lists from the same request, and one of them built nothing usable at all.
What makes this different from just being invisible
We've written before about categories going invisible across the board. Med spas in Los Angeles came back zero for zero across both models. This is a different kind of failure. GPT-4o managed a partial answer here, three names, not impressive, but a real answer with actual businesses in it. Perplexity returned nothing usable for the category at all.
That gap points to something specific about how each model constructs an answer. GPT-4o is willing to pull from mixed sources and name a business even with thinner backing. Perplexity relies more heavily on live retrieval, and if it can't locate a source it trusts enough to quote directly, it hedges rather than guesses. Boutique salons lack an equivalent of RealSelf or Healthgrades, the kind of niche authority site that gives an AI model something concrete to cite. Med spas at least get swept into aesthetics and dermatology content. A neighborhood hair salon competing on style and personality has no obvious category of source material backing it up anywhere online, and Perplexity's caution shows exactly where that gap lives.
Why this matters more than a single-engine miss
If you're a boutique salon owner and you've heard "we're doing fine on ChatGPT," that sentence is doing a lot of unearned work. Fine on ChatGPT and invisible on Perplexity is a real, common outcome, not some edge case. Third-party estimates put Perplexity past 100 million monthly users sometime in 2026, up from a fraction of that a year earlier. Perplexity doesn't publish its own numbers, so treat the exact figure loosely, but the trajectory isn't disputed: it's a genuinely large and fast-growing slice of people who ask a question and expect a named answer, not a page of links to click through. A business that only appears for half the AI traffic asking about it is leaving the other half to whoever happens to have better source coverage, not necessarily whoever runs the better salon.
The three salons that made GPT's list share something the other seven didn't: each one shows up in more than just Yelp and Google. Avant-Garde won a Gold "Best Hair Salon" title in the Miami Herald's Best of Miami-Dade and landed coverage in Allure and Miami New Times. Oren Salon appears in Dish Miami's hair salon guide and got an owner interview in Voyage MIA. Danny Jelaca's editorial work has run in Vogue, Harper's Bazaar, and Ocean Drive, plus an InStyle "Best Cut" title. None of that is a backlink to their websites. It's the salon's name sitting in someone else's published sentence, exactly the kind of evidence a model can point to instead of taking the business's own word for it.
What to actually check this week
Search your salon's name in quotes and see what surfaces beyond your own site, Google Business Profile, and Yelp. If the honest answer is "nothing," that's the real problem, not your website copy or your Instagram grid.
Look for a local "best salons in [your city]" piece from an actual publication, not some listicle site that exists purely to rank in Google. If one doesn't exist yet, plenty of local outlets will bite on a genuinely good pitch, especially around a new stylist, a seasonal color trend, or a charity event. That's exactly the kind of third-party mention Perplexity leans on when deciding whether to name someone.
Check whether any press mentions or roundups exist for your stylists individually, not just the salon as a brand. AI models treat named experts differently than they treat a storefront, and stylist-level coverage reads as a stronger trust signal than a business listing.
Then run the actual check. Type your salon's name and city into ChatGPT and Perplexity separately and see what each one returns. Don't assume a good result on one means you're covered on both, because in this data, that assumption was wrong nine times out of ten.