Automatic Session Analysis

When a session recording is complete and the MP4 video is available, Revisit automatically:

1. Uploads the MP4: Sends the session video to Google Gemini for analysis
2. Applies Context: Uses your project type and interests to tailor insights
3. Generates Analysis: Creates a comprehensive report with multiple components
4. Stores Results: Saves the analysis in your project's database for future reference

Analysis Components

Each AI analysis includes several key components:

• Title: A concise, compelling description of the user's journey
• Summary: 2-5 sentences about user intent, behavior, and key insights
• Insights: Actionable bullet points highlighting friction points, motivations, and UX issues
• Timeline: Annotated time ranges (MM:SS) covering the entire video duration
• Raw Output: Complete JSON payload for traceability and debugging

Context and Customization

AI analysis is tailored to your specific project needs:

• Project Type: Analysis focuses on relevant KPIs (e-commerce checkout, SaaS signup flows, etc.)
• Interests: Prioritizes findings related to your selected focus areas
• Privacy Awareness: Recognizes masked content and doesn't speculate about redacted information
• Issue Creation: Can automatically create GitHub issues or Jira tickets when problems are detected

Chat Interface

Access AI chat through multiple interfaces:

• Session-Level Chat: Ask questions about specific sessions in the session view
• Project Chat: General AI chat about your project's session data
• Chat History: All conversations are saved and can be revisited
• Quick Actions: Pre-built questions for common analysis needs

How Chat Works

The AI chat system uses several data sources to provide accurate answers:

• Session Video: Direct access to MP4 recordings for ground-truth answers
• Analysis Context: Latest AI analysis including title, summary, insights, and timeline
• Chat History: Previous messages in the conversation for context
• Struggle Detection: Automated detection of user frustration patterns
• Session Links: Can reference specific sessions and timestamps

Streaming Responses

AI responses are delivered in real-time using streaming technology:

• Token Streaming: Responses appear as they're generated, not all at once
• Real-time Updates: See the AI "thinking" and responding in real-time
• Session Links: AI can insert clickable links to relevant sessions
• Error Handling: Graceful fallback if streaming encounters issues
• Auto-scroll: Chat automatically scrolls to show new content

Advanced Chat Features

The chat system includes several advanced capabilities:

• Conversation Management: Create, switch between, and delete chat conversations
• Auto-titling: AI generates descriptive titles for conversations
• Session References: Click links to jump directly to relevant session moments
• Context Preservation: Chat history maintains context across messages
• Quick Questions: Pre-built prompts for common analysis scenarios

Asking Effective Questions

Frame your questions for better AI responses:

• Be Specific: "Why did users abandon checkout?" vs "What's wrong?"
• Include Context: Mention specific pages, features, or user segments
• Ask Follow-ups: Build on previous answers for deeper insights
• Use Examples: Reference specific sessions or timeframes when relevant
• Focus on Actions: Ask what you can do to improve the experience

Common Use Cases

AI analysis and chat are particularly useful for:

• Behavioral Analysis: Understanding user motivations and pain points
• Technical Issues: Identifying JavaScript errors, slow loading, broken links
• Conversion Optimization: Finding barriers to successful conversions
• UX Improvements: Discovering usability issues and friction points
• Performance Monitoring: Correlating technical issues with user behavior
• Mobile vs Desktop: Comparing experiences across devices