Reputation

Open Reports → Reputation to review your brand and tracked competitors against the dimensions configured for the project. Reputation reports use already collected responses; they do not replace running your prompts.

Prepare and generate a reportCopy Anchor Link

1. Collect answers to questions relevant to the attributes you want to investigate.
2. Open Configure dimensions and review the dimensions before saving changes.
3. Select Generate reputation report when no report is open, or Regenerate to create a new report from available evidence.
4. Use All reports to return to the saved report library.

Generating or regenerating is a write action. Changing dimensions and generating reports should be deliberate, not a way to refresh a read-only view. A dimension needs relevant prompt or answer evidence; adding its name does not manufacture supporting observations.

Read OverviewCopy Anchor Link

Set the date range and All filters, then inspect Brand comparison, AI reputation score, and Dimension breakdown.

• The overall reputation score averages dimensions that have evidence. Unscored dimensions are excluded, not treated as zero.
• Standing is relative to brands with evidence in the current tracked set.
• Mention share compares tracked brand mentions, not the entire market.
• Dimension coverage distinguishes relevant evidence for a specific attribute from a broad overall signal.
• Visibility and citation authority are indexed relative to the strongest tracked brand. Sentiment uses an absolute scale with 50 representing neutral.

A leader-relative score can change when competitor evidence changes, even without a corresponding change in your own responses. Read response counts, evidence tiers, and reasoning alongside the score.

Inspect model and raw evidenceCopy Anchor Link

Open Model Comparison to compare dimension scores by model and focus on the brand or dimension offered by the controls.

Open Raw Data for rows by brand, model, and dimension. Use its dimension filter and search, and disable Hide unscored when you need to see gaps rather than only supported scores. Read the evidence tier and reasoning before interpreting a number as a robust conclusion.

Use Export CSV for the relevant report or raw-data view. Keep the selected scope with the export when sharing a comparison.

TroubleshootingCopy Anchor Link

• Generate is unavailable: collect tracked AI responses first.
• Responses exist but there is nothing to score: the selected answers may not mention your project or tracked competitors.
• A dimension has no score: inspect whether relevant questions and answers support it; unscored is not poor performance.
• No rows match: widen the date range or clear filters.
• Scores differ from Overview: these are dimension/evidence and competitor-indexed calculations, not the Overview visibility average.

Related guidesCopy Anchor Link

Prompts · Sentiment · Competitors