Every practice session in Oliver AI generates a detailed call record. These records form the foundation for evaluations, analytics, and coaching recommendations.
When you complete a practice session, the following data is captured and stored:
- Full transcript -- Every message exchanged between you and the AI customer persona
- Call duration -- The length of the practice session
- Call type -- The scenario category (cold call, discovery, demo, etc.)
- Product context -- Which product or service was being discussed
- Customer persona -- Details about the AI customer you interacted with
- Summary -- An AI-generated summary of the call highlighting key moments
- AI usage -- Usage recorded for the AI and voice services behind the session
- Timestamp -- The exact date and time of the session
After a practice session ends, Oliver AI automatically generates a call summary using AI analysis. The summary captures:
- The main topics discussed during the call
- Key objections raised by the customer and how they were handled
- Any commitments or next steps agreed upon
- Notable strengths or areas for improvement
The summary is generated using the same AI models that power the practice sessions, with extraction prompts designed to identify the agent's name from the transcript and highlight the most important moments.
- Navigate to Practice History (under the My Progress section in the sidebar).
- You will see a list of your past practice sessions sorted by date.
- Each entry displays:
- Customer name (from the AI persona)
- Product discussed
- Call type and duration
- Date of the session
- Click on any call to view the full transcript and summary.
Managers have additional visibility:
- View call records for all team members in your organization.
- Filter calls by team member, product, call type, or date range.
- Access aggregated call statistics across the team.
- View call data alongside evaluations for comprehensive performance review.
Each call record includes these fields:
| Field | Description |
|---|
| ID | Unique identifier for the call |
| User ID | The sales rep who made the call |
| Product ID | The product discussed |
| Call Type | Scenario type (cold_call, discovery, demo, etc.) |
| Transcript | Full text of the conversation |
| Summary | AI-generated call summary |
| Duration | Length of the call in seconds |
| Customer Name | Name of the AI customer persona |
| Target Audience | The customer segment being targeted |
| Created At | Timestamp of when the call occurred |
Some calls include extended data that provides deeper analysis:
- Sentiment analysis -- How the conversation tone shifted throughout the call
- Key phrases -- Important terms and topics identified in the transcript
- Engagement metrics -- Measures of how well the conversation flowed
Note: Analysis capabilities may vary -- some features listed above are planned for future releases. The availability of extended data depends on your organization's configuration and the specific analysis features currently enabled.
- All call records are stored permanently in your organization's database.
- Transcripts and summaries are available for review at any time.
- Call data is used to calculate analytics and leaderboard rankings.
Note: Call records are scoped to your organization. Sales reps can only see their own calls, while managers can see calls from their team members. Admins and superadmins have organization-wide visibility.
Review your call history regularly to identify patterns:
- Compare transcripts -- Look at how your approach changes across sessions.
- Track summary trends -- See if the same objections keep coming up.
- Check durations -- Very short calls may indicate you are not engaging deeply enough. Very long calls may suggest you need to be more concise.
- Cross-reference with evaluations -- Pair your call history with evaluation scores to understand which approaches lead to better outcomes.