What AI answer engines cite from a medtech site
Thirteen questions put to ChatGPT, Claude and Grok, five separate times each, with both outcomes counted separately: whether the engine read a page, and whether it named the brand. 60.0% of 195 calls read a page and 45.1% named us, but half the question and engine pairs that named us at all did not do it on every run. Includes our own 1,158 pages scored against the same standard, published with the criteria we fail.
Read the study →