What are Daily Insights?
Daily Insights are AI-generated observations about your user sessions. Every night, Rehearsals analyzes your new session data to surface patterns, anomalies, and interesting behaviors that you might have missed. These appear as cards in the right-hand sidebar of your Sessions page.How It Works
Data Collection
Throughout the day, Rehearsals collects session replays, click maps, and user journey data from your website.
Nightly Analysis
At midnight EST, our AI reviews the day’s sessions, looking for:
- Friction Points — Rage clicks, rapid scrolling, or repeated errors
- Conversion Blockers — Why users dropped off at specific steps
- Success Patterns — What successful users did differently
- Anomalies — Unusual spikes in traffic or specific behaviors
Insight Generation
The system generates 3–5 actionable insights based on these findings and queues them for your review.
Requirements
To ensure high-quality insights, generation only triggers when:- Your project has 2,000+ total sessions recorded
- You have fewer than 3 unacknowledged insights in your queue
If you stop seeing new insights, check if you have pending cards in your sidebar. Vote on them (thumbs up or down) to clear the queue and resume generation.
Interacting with Insights
Each insight card offers several actions:| Action | What It Does |
|---|---|
| Ask AI | Opens a chat with the Agent, pre-loaded with context about the insight. Ask follow-up questions like “Show me the sessions that caused this.” |
| Thumbs Up | Marks the insight as valuable. Helps train the AI to find similar patterns in the future. |
| Thumbs Down | Dismisses the insight as irrelevant. The AI learns to deprioritize this type of finding. |
Best Practices
Review Daily
Make it a habit to check your insights each morning. Patterns are easier to act on when they’re fresh.
Validate with Replays
Use “Ask AI” to pull up the specific sessions that triggered an insight. Watching actual user behavior is the best way to validate the AI’s observation.
Clear Your Queue
Vote on cards consistently to keep new insights flowing. A backed-up queue pauses generation.
Look for Trends
Similar insights appearing multiple days in a row often indicate a systemic issue worth prioritizing.