Same brand, different ChatGPT prompt, totally different rankings

I've been testing whether AI visibility matches Google rankings. Short answer: not even close, and the exact prompt matters more than I expected.

Incognito, same day, ChatGPT:

Prompt 1: "best CRM for small business"
HubSpot came up first.

Prompt 2: "affordable CRM for startups"
Different list:

1. Attio
2. Folk
3. HubSpot (startup program only, with caveats)
4. Zoho Bigin

HubSpot isn't invisible. But they're not winning the prompt a cashstrapped founder would actually type.

Attio and Folk get top spots because the answer focuses on price and founder-led sales. HubSpot's slot is basically "apply for a discount if you're in an accelerator."

If you're only checking one generic prompt you might think you're fine when you're losing on buyer-intent queries.

I'm building CiteLens to track multiple buyer prompts weekly because one check lied to me. Early stage, trying to see if the output is useful before I keep building.

Anyone else running these tests? What prompts are you tracking?