57% of US local markets had a different #1 AI recommendation this week.

we asked ChatGPT, Perplexity, Claude, and Gemini 41,400 real customer

questions like "best hotels in New York" and "top HVAC in Phoenix" across

690 US markets. Then I asked the exact same questions 7 days later.

391 of 690 markets had a completely different #1 answer.

Volatility by industry:

▸ HVAC contractors: 69% of markets flipped #1

▸ Motels: 66%

▸ Personal-injury lawyers: 64.5%

▸ Dentists: 60%

▸ Hotels: 49.5%

▸ Med spas: 45%

▸ Wedding venues: 43%

Not one industry stayed below 40%.

Named-brand examples that were #1 last week and GONE this week:

▸ InterContinental Miami — was #1 with 20 citations, zero now

▸ Hotel Nikko San Francisco — from #1 to #15

▸ The Tinsmith (Madison, WI) wedding venue — from #1 to #27

▸ Cooper Hurley Injury Lawyers (Virginia Beach) — from #1 to #18

None of these businesses closed. The AI just started naming other names.

If you're a local business owner: getting cited by ChatGPT ONCE ≠ being

cited next week. Multi-engine coverage matters way more than peak position in one engine.

12 views

Add a comment

Replies

Best

Interesting data.

I think it also shows how early AI search still is. Rankings seem much more volatile than traditional SEO, so relying on a single AI engine probably isn't a good strategy. for local businesses, consistent visibility across multiple platforms feels more valuable than being #1 in one engine for a short time.

I think the bigger issue is trust. If the same model gives different recommendations to the exact same questions just a week later, how do users know which answer is actually better? And as AI search grows, I wonder how these systems will ensure recommendation quality while remaining resilient to manipulation through artificial authority signals rather than genuine quality 🤔