Responses

Responses are the captured outputs from AI models after Cleotic runs your prompts. They are the source data behind visibility, sentiment, answer-gap, and citation analytics.

What a response contains

Every response can include:

Response text

The full model output is available in two views:

• Parsed: a formatted reading view
• Raw: the original text payload

Brand mentions

Cleotic detects which configured brands were mentioned and classifies each mention as:

• Positive
• Negative
• Neutral

Mentions are shown as badges alongside the response rather than as inline highlights in the body text.

Citations

If the model referenced sources, Cleotic captures the cited URLs and associated metadata so they can flow into project-level citation analysis.

AI Insight

Some responses include an AI Insight card that summarizes things like inaccurate claims, concerns, wins, and suggested actions.

Viewing responses

From a prompt's responses page, you can inspect outputs by model and capture time. This is useful when you need to understand why a visibility score moved or why a brand was classified a certain way.

If a trial has moved into its restricted state, older response content can be locked behind an upgrade prompt.

How responses feed analytics

Response data is aggregated into:

• Visibility scores
• Share of voice
• Sentiment analysis
• Citations
• Answer gaps

Tips

• Check raw responses when a metric changes unexpectedly.
• Reanalyse after updating brand aliases so historical responses can be reprocessed.
• Compare the same prompt across models to see where perception differs.