Model and prompt breakdowns
Model and prompt breakdowns reading
Breakdowns show where the overall number comes from. Use them to see which model or prompt changes the story.
This prevents one score from hiding the useful detail.
Model-specific movement
Different AI systems can name different brands. If one model changes, inspect its sources and answer wording.
For example, ChatGPT may name you because it reads your site, while Gemini may name a competitor because it reads review pages. That is not random noise. It is a clue.
Comparison use
Compare models only after checking that the same or similar prompts were run.
Prompt mix matters.
Breakdowns
Model and prompt breakdowns show where visibility changes come from instead of hiding them in one number.
| Breakdown | Useful question |
|---|---|
| By model | Does one AI system understand us better? |
| By prompt | Which buyer question creates the gap? |
| By source | What did the model read before answering? |
| By competitor | Who gained the answer space? |
Evidence follow-up
Ask CLEA which model or prompt deserves review first. Then inspect that run, not the whole dashboard.
prompt.txtCopy prompt
Find the model or prompt that explains this visibility change. Keep the answer short, and point me to the exact evidence to inspect.