AI Search Notes
Research, experiments and guides on what gets a brand cited in AI search. Every stat checked.
-
How to prove your AI search work is paying off
One visibility score can make you look better than you are. Here’s how I’d set up AI search tracking so you can show your CFO which buyers you’re actually winning.

-
Businesses in ChatGPT: what marketers need to know
OpenAI now lets businesses claim their profile inside ChatGPT. What it does, what it won’t, what our data says about getting in early, and what I’d do this week.

-
Turn up the prompt volume
At 10 prompts, a brand’s AI visibility score moves about 4x as much day to day as at 1,000, even when nothing changed. More prompts give you a number you…

-
Copilot is the strictest buyer in AI search
Copilot says something bad about a brand about 3x as often as ChatGPT. Price is where it’s toughest, and most brands aren’t even tracking it.

-
How many shoppers use AI right before they check out?
About 1 in 5 US checkouts now has AI use in the 24 hours before, up from 16% in June. Google AI Mode drove most of the late-August jump.

-
Test your brand in ChatGPT with 50 prompts
The 50 prompts from my LinkedIn post, as an experiment you can run on your own brand. How to set it up, what to log, and what your results mean.

-
Which sources do AI engines cite for local businesses?
ChatGPT cites .gov licensing pages for local searches. Claude cites Yelp. Here’s the data on both, and what local businesses should do about it.

-
Does being named first keep a brand in ChatGPT’s answers?
Brands ChatGPT names first are still named for the same question 7 weeks later 79% of the time. Brands named 6th to 10th, only 37%. Here’s the full data and…

-
What to focus on in AI search, in order
The order I’d go in with AI search: six steps, from saying what you sell in plain words to getting cited on the sites AI already trusts, with the data…

-
Which AI models should your brand actually track?
Start with ChatGPT, then add the 2 or 3 AI models your buyers actually use. Here is who is on each model and the research behind every pick.

-
The content freshness stats everyone shares, checked
I traced the AI search freshness stats everyone repeats back to their sources. Two hold up, three I couldn’t verify, and two get stretched.

-
The pages that get cited most in ChatGPT aren’t the ones that last
Guides and lists win ChatGPT citations fast, then fade. Reference and research pages barely move. 7 weeks of data by page type.

-
Why I’m testing AI search on my own name
I’ve been in AI search since it basically started. This site is where I actually do it, starting with my own name.

