Generative engine optimization · Citation growth
ChatGPT citations are up 4x this year. In some local categories, directories still take almost half.
A report that dropped in May tracked how often ChatGPT attaches a source link to its answers, month by month, across US prompts. Back in June 2025 that happened roughly 1.6% of the time. By May 2026 it had hit about 6.8%, more than four times higher in less than a year.
The climb wasn't steady. It spiked hard starting October 2025, flattened out for around four months, then accelerated again heading into spring. But the trajectory holds: citations are becoming standard in more ChatGPT answers instead of staying an occasional footnote.
The growth isn't even
Here's what matters more than that headline figure. The growth isn't spread evenly across categories. Travel and hospitality prompts get a citation about 23% of the time, the highest of any category named in the report. Automotive sits around 20%. Professional services, the report's bottom category, gets cited under 4% of the time, barely above where the entire platform stood a year ago.
The report's explanation is straightforward. Categories where people compare specific, verifiable options, one hotel against another, one car against another, get cited heavily. Categories built on general advice, with nothing concrete to point at, lag way behind.
That distinction ought to be encouraging if you run a local service business. "Best med spa in Miami" isn't general advice. It's a comparison between named, checkable places, structurally closer to choosing a hotel than to asking for career guidance. By the report's own reasoning, local service categories should sit up near travel and automotive rather than down with professional services.
What we checked
The report doesn't say where local services actually fall. It names three categories and stops there. So we checked what we already had on hand.
Through August we logged every citation ChatGPT and Perplexity produced across ten local-service categories, five cities each, 1,423 citations total: dentists, gyms, plumbers, hair salons, med spas, optometrists, auto repair shops, chiropractors, personal injury lawyers, and veterinarians. For each vertical we counted average sources per query, and how many of those sources were recognized directories and review aggregators rather than an individual business's own site.
Dentists led the pack on volume, averaging 7.9 sources per query. Gyms and plumbers followed tight behind, both above 7.5. Veterinarians brought up the rear at 5.9. Every category sat well above what a sub-4% citation rate like professional services would ever generate. Local service businesses do get compared, named, and cited at volumes that fit the report's checkable-options theory.
Volume isn't the same as landing on the business
But volume doesn't mean the citation reaches the business. Here's how much of each category's citations went to directories instead of individual sites, from worst to best. Auto repair gives up 43.2% of its citations to directories, the highest of the ten. Optometrist loses 36.6%. Hair salon loses 34.9%. Gym gives away 34.4%. Med spa loses 32.4%. Dentist gives up 31.0%. Veterinarian loses 27.1%. Chiropractor gives up 26.3%. Plumber gives away 26.1%. Personal injury lawyer does best at 24.2%, and still loses nearly a quarter of its citation activity to legal directories and referral sites.
Notice what doesn't track. Dentist generates the most citations per query of any category we tracked, yet still routes about a third to directories. Veterinarian generates the fewest citations per query, yet keeps a slightly better share on individual clinics than several busier categories. More citation activity in a category doesn't promise more of it reaches the actual business. Sometimes it just means more traffic through the same directory gates.
We should be straight about what these two numbers can and can't be woven together into. The report's category breakdown runs through May 2026 and doesn't reveal sample size. Our data came from August 26 and 27, a few months later, on a narrower slice: ten specific local-service categories rather than the report's broad commercial sectors. We're not claiming the two datasets form one continuous thread. What we can say is they point the same direction from two different vantage points: the citation pie for checkable categories keeps expanding, and a meaningful slice of it, sometimes approaching half, isn't landing on the business being compared. It's landing on whatever sits between the customer and that comparison.
What this changes
That shifts what the fix looks like. If your category were stuck under 4% like professional services, the battle would be getting cited at all. That's not the local-service situation. ChatGPT already cites something in nearly every one of these searches, seven sources per query on average. The real fight is over which seven get selected. In every category we checked, at least a quarter of those slots go to a directory instead of you.
The businesses that appear as themselves rather than a line inside someone else's aggregator tend to share one trait: pages built around the specific service a customer asked for instead of a generic services list, plus a review and listing footprint robust enough that ChatGPT doesn't need to lean on RealSelf or Yelp to fill the gap.
A free check shows whether your business currently sits on either side of that line.
Source: Similarweb, "AI Search Stats 2026" — US ChatGPT citation rate, June 2025 to May 2026, category breakdown for travel & hospitality, automotive, and professional services (industry research, methodology and sample size not disclosed). Our own figures come from askchat.studio's GEO probe database: 1,423 citations across 10 local-service categories and 5 cities each, collected August 26-27, 2026.