The comparison decision
Use a common prompt panel, platform scope, market, date range, and missing-data policy for every competitor.
Evidence-led comparison method
- Define competitors and aliases before extraction
- Collect successful answers and preserve failed or unavailable observations separately
- Measure coverage and position within each Prompt Group
- Review Sentiment, product descriptions, and citations at the snapshot level
- Map repeated competitor advantages to pages, entities, evidence, or external sources
Competitor example
A competitor may have fewer mentions but appear in high-intent comparison prompts with authoritative citations. That pattern can matter more than a larger total generated by generic prompts.
Evidence to retain for competitive analysis with generative AI search data
Keep these fields with the decision:
- prompt
- market
- platform
- competitor
- answer evidence
- cited URL
- observed difference
- hypothesis
- owner
- retest date
Limits and UnderAI's role in competitive analysis with generative AI search data
AI answers do not reveal the ranking algorithm or prove why a competitor appeared. Treat source and page gaps as hypotheses to test.
UnderAI combines configured competitor tracking, Platform Overview metrics, Answer Snapshots, and citation records within the same project scope.