Gather Insights
Analyze assistant sessions and extract valuable data from conversations
The Gather Insights page (also known as the Analysis tab) allows you to enable automatic analysis of your assistant’s conversations.
When analysis is enabled, Ziddny reviews each completed session and generates insights based on your custom analysis prompt — such as summarizing feedback or extracting user data.
Enable Analysis

To start using this feature:
- Check Enable Analysis to activate post-session analysis.
- After each session, Ziddny automatically runs the analysis against the session’s transcription.
- The results will appear later under your assistant’s Metrics page.
This helps you gain structured insights, such as user satisfaction, lead conversion signals, or interview outcomes.
Analysis Prompt
The Analysis Prompt is where you define what type of insights the system should generate.
You can write short, clear instructions describing what the analysis should focus on.
Examples:
- “Summarize the user’s feedback and sentiment.”
- “Evaluate how well the assistant handled the user’s questions.”
- “Extract key details from the conversation for sales opportunities.”
- “Score the user’s interview performance from 1 to 10 based on confidence and clarity.”
Keep the analysis prompt simple and focused for best results.
Data to Extract
You can also specify structured data fields to extract from each conversation.
For each field, define:
- Name – The attribute identifier (e.g.,
user_name,age,role,experience) - Description – A short explanation of what the data represents
Example: “The name of the user” or “The user’s job title”
Click Add to include multiple data points.
The assistant will automatically extract this information from each conversation and make it available for reporting.
Viewing Analysis Results
All analysis results and extracted data are displayed under the Metrics page for your assistant.
There, you can:
- Review insights and summaries generated from sessions
- Track structured attributes (e.g., user names, roles, satisfaction levels)
Example Use Cases
- Customer Feedback: Summarize satisfaction levels or highlight common issues.
- Interview Evaluations: Score candidates and extract their background information.
- Sales Analysis: Detect lead quality or intent from chat transcripts.
- Education or Training: Track progress and extract student performance data.
Next Steps
Once you’ve enabled and configured analysis:
- Test your assistant and complete a few sessions.
- Visit the Metrics page to see the collected insights.
- Adjust your analysis prompt or extracted fields based on what you want to track.