Blog
Blog
Notes on how we measure AI recommendations, and what we learn from it.
2026
- AI answers change every time you ask A 2026 study found AI assistants almost never give the same list of brands twice. What it says, what it means for measuring, and where it falls short.
- API answers and app answers are not the same Asking an AI through its API is not the same as typing into its app. What one 2026 study measured, how we label the gap, and what it means for you.
- Blocking AI bots: what you gain and what you lose Training bots, search bots and user-request fetchers are different things. A decision table for small businesses, based on what the companies document.
- Does llms.txt do anything for AI search? What Google says about llms.txt, what a 137,000-site log study found, and the few cases where a small business might still want one.
- SEO, AEO and GEO in plain English Three acronyms for one worry: being found when people ask machines. What each means, where they overlap, and what changes for a small business.
- Why we ask every question twice AI answers change from one ask to the next. Here is how that shapes the way we measure, and why we show a range instead of one number.