Exploring Findings

Finding Types

• Insight: key information or pattern worth tracking.
• Risk: threats or blockers that could affect outcomes.
• Opportunity: favorable conditions or openings to pursue.
• Recommendation: actionable next steps derived from evidence.

Filtering and Grouping

• Search across content, reasoning, and artifact columns.
• Filter to a specific finding type or view all types.
• Group by type or source artifact to reveal recurring patterns.
• Switch between card and table views based on how you want to review data.

Human Expertise: Grounding in Gold Standard

AI extracts findings quickly, but accuracy comes from human judgment. When you review, edit, and approve findings, you create a "Gold Standard" — a verified foundation the AI uses for all future syntheses and decisions. You stay in control; the AI stays grounded.

• Mark findings as reviewed. Use Mark as reviewed in the action menu to flag that human expertise has validated a finding. Reviewed findings display a Reviewed badge.
• Edit to verify. When you edit a finding inline, you automatically mark it as reviewed — because a human was involved in producing the final version.
• Weight during synthesis. During synthesis, reviewed findings are prioritized with higher confidence. The model is explicitly instructed to draw conclusions from your verified evidence, making strategic outputs far more reliable.

Build your Gold Standard over time — each verified finding strengthens future syntheses and becomes a reusable asset across all projects that cite this evidence.

Export and Actions

• Export findings to Markdown with grouped sections, source links, and reference snippets when available.
• Row actions let you edit in context, open the source artifact, or jump directly to cited lines.
• Column visibility, sorting, grouping, and view preferences persist between sessions.