I help founders & SEOs get cited by ChatGPT, Perplexity & AI Overviews. Daily teardowns of who AI quotes & why. AI replaces thinking, not understanding.
I analyzed how 3 AI engines actually pick what to cite. They could not be more different — and "optimize for all three the same way" is why most GEO is flat. Breakdown:
Want me to run this for your category? Drop your niche and the buyer query your customers would type, and I'll screenshot who ChatGPT actually cites. First 5, free.
I asked ChatGPT for the best project management software for marketing agencies. It thought for 3m 13s.
The winner wasn't Asana. Wasn't https://t.co/cXcbRLncUk. Wasn't ClickUp.
It picked https://t.co/2kN0WyQPjT at #1. The brand most people have never heard of.
Full teardown 👇
The lesson: you don't win AI recommendations by being the biggest brand. You win by speaking your exact buyer's language in specific, citable text on your own pages.
If an AI ranked your category tomorrow, would it quote your page or your competitor's?
ChatGPT's own close: trial Teamwork, Productive and https://t.co/cXcbRLncUk on one real client campaign. Not the famous names by default, the specialists that match the work.
That's the shift. Fit beats fame.
The numbers that survived were all from official pages: Productive at $10/user/mo with a 3-seat min, monday Pro's 25,000 actions/month, Teamwork's free 5 users / 5 projects.
Quotable specifics on your own pages are what ChatGPT repeats back.
Now the part that should change how you write pages. Every source ChatGPT cited was the vendor's OWN site. Teamwork, Productive, Wrike, monday, Asana, ClickUp, Scoro.
Zero affiliate listicles. Zero "top 10" roundups.
ChatGPT even flagged the gap on Asana directly: cleaner than ClickUp, but agency profitability, invoicing and resource planning need add-ons.
That's the tell. The model isn't ranking by fame. It's ranking by fit to the persona in the query.
Why did the smaller name win? https://t.co/2kN0WyQPjT is agency-native. One platform for projects, clients, time tracking, proofs, utilization, budgets, retainers, invoicing and profitability. Plus a free plan up to 5 users / 5 projects.
It was built for the exact buyer asking.
Here's the full ranking ChatGPT gave:
1. https://t.co/2kN0WyQPjT
2. Productive
3. Wrike
4. https://t.co/cXcbRLncUk
5. Asana
6. ClickUp
7. Scoro
Asana and ClickUp, two of the most famous names here, got buried at #5 and #6. Brand awareness did nothing for them.
Why did the smaller name win? https://t.co/2kN0WyQPjT is agency-native. One platform for projects, clients, time tracking, proofs, utilization, budgets, retainers, invoicing and profitability. Plus a free plan up to 5 users / 5 projects.
It was built for the exact buyer asking.
Unpopular take: stop selling llms.txt as your AI shortcut. No major AI engine has confirmed it reads the file, and I've never seen one citation move from it. What moves citations is being the clearest answer on a page engines already crawl. Seen llms.txt change your visibility?
@Kevin_Indig Right? Though the recovery I'd really watch is their AI citations, not Google. For a hands-on testing site like HouseFresh those two can fully decouple. In your citation data, do they move together, or split?