First test pending
This page fills in from the first consistency test. Until then there is nothing to show, and we do not fill the gap with estimates. The other case studies below already carry numbers.
Engines
ChatGPT, Claude, Google AI Overviews and Gemini
Runs
Every prompt asked 3 times on every engine
Labels
Strong 3 of 3, Moderate 2 of 3, Weak 1 of 3, Invisible 0 of 3
Fix log
Opens with the first fix, dated, beside the next test