PondralCorrection, August 2026: the study described below was never run. This post said we queried five engines three times a day for ninety days, which would be 1,350 measurements. It did not happen. Our production database holds no measurement of any kind before this post was published on 14 March 2026 — no scored results, no audit runs, and no API calls to any engine. The first data we hold dates from 2 May 2026, seven weeks later. Nothing below rests on a measurement, and the conclusion it reports should not be read as one of ours.
We are leaving the post up rather than deleting it, because a quietly removed page is not a correction. What follows is the original text.
We ran the same query, "best CRM for a 50-person remote team," through ChatGPT, Claude, Gemini, Perplexity, and Grok, three times each, every day for ninety days. For context on how AI engines differ in how they cite brands, see our ChatGPT vs Perplexity vs Gemini comparison.
Across the runs, no single brand made every top-three list. A handful of incumbents dominated, but the order shifted constantly. We saw a similar pattern in our broader research on which brands dominate AI answers and why.
The criteria the models cited also shifted over time. The dominant ranking criterion changed week to week: "ease of use," then "AI features," then specific integrations. The underlying products barely moved. Our scoring rubric explainer breaks down how we measure these shifts systematically.
AI visibility differs from SEO here: the answer changes when the model's framing changes, not only when your content does. You can do everything right and still drop because the model decided the question is now about something else. The metric that matters is AI Visibility, not branded search volume. Pondral's complete guide to AI visibility covers how to measure and improve it.