Help › Understanding your data

Why the assistants disagree with each other

Different sources, different training, different recency. Disagreement is information.

Last updated 2026-07-30

It is common to be strong in Perplexity and invisible in ChatGPT, or the reverse. This is not an error in the measurement.

Assistants differ in whether they search the live web at answer time, which sources they favour, how recent their knowledge is, and how willing they are to name specific companies at all.

How to read the split

• Strong in search-grounded assistants, weak elsewhere: your presence is recent or lives on the live web. It has not settled into model knowledge yet.
• Weak in search-grounded assistants specifically: the sites being retrieved for your category do not cover you. Go to Sources.
• Weak everywhere: this is a coverage problem, not an engine problem.

Which engines you see

Your plan decides how many assistants we check and, on Starter and Agency, which ones you choose. The model filter on Overview only offers engines your plan actually covers.

Related

• Visibility, Authority and Position — what each one means
• Why other people's websites decide your AI visibility
• Why your sentiment score is probably 50
• Why your score moved (or didn't)