Troubleshooting

The report contains fewer prompts than the plan maximum

The Prompt Set uses qualified, relevant, and sufficiently diverse candidates. If too few qualified unique prompts are available, SpyderBot may use a smaller Prompt Set rather than adding filler prompts.

Executions are higher than the number of prompts

Prompts may be executed on several AI systems and may be repeated under the run policy. Executions are prompt-to-system attempts, not unique prompts.

Prompt-system Coverage is complete but response counts differ

Complete cell coverage means each eligible prompt-system pair has at least one valid response. Cells may still contain different response depth.

A metric is unavailable

Common reasons include no eligible responses, missing denominator, insufficient sample, incompatible report scopes, AI-system unavailability, failed validation, restricted evidence, or required tracking not connected. Unavailable data must not be interpreted as zero.

Report totals do not match an older report

Check Prompt Set, AI-system scope, competitor set, geography and language, valid-response population, metric-definition version, extraction version, and Stable matched subset.

Sentiment does not match the website's intended messaging

Brand Insights reports how selected AI systems represented the brand in collected responses. Investigate through Evidence, Prompt Intelligence, or GEO Audit.

Founder Negative Context appears high

Confirm Organic versus Risk-aided population, founder entity resolution, sample count, current versus historical context, allegation versus verified fact, and supporting responses. Do not treat the percentage as a corporate risk score.

Trend is unavailable

Trend is suppressed when report scopes or methodology versions are not sufficiently compatible. Use the displayed compatibility reason and compare the Stable matched subset where available.