Results: How many calls did the agent actually process?
In this module, you will find the results of all completed calls, allowing you to evaluate the agent’s performance from a macro perspective:
- Total number of calls processed.
- Distribution by result type (answered, failed, IVR, unanswered, duplicated, busy, postponed by offset - outside business hours).
- The agent’s real effectiveness within the stack.
This analysis is key to validating the agent configuration, retry logic, and message quality.
From data package to decision-making:
Results export
The implementation of call scoring and feedback provider in Rootlenses Voice enables:
- Increase engagement with more human and relevant conversations.
- Reduce hang-ups and early friction.
- Quickly detect data, timing, or messaging issues.
- Iteratively improve agents, voices, and COTs with real evidence.
- Turn every call into learning for the next one.
The value is not only in viewing the data, but in turning it into actionable insights.
Call detail
Each call includes a detailed view that consolidates all critical information:
Call information
- Customer name.
- Dialed phone number.
- Line status.
- Call result (successful, failed, etc.).
- Processing start and end date/time.
- Call duration.
This allows you to reconstruct exactly what happened and when it happened.
Automatic call analysis
Rootlenses Voice applies intelligent analysis to each interaction:
- Interest level: Score generated based on the real interaction with the agent.
- Suggested follow-up method.
- Identified contact data (email, phone, contact).
- Call termination: Indicates whether the user or the agent hung up.
- Call recording: to evaluate the AI’s interaction with the user.
This analysis enables prioritizing efforts, focusing follow-ups, and avoiding intuition-based decisions.
Call recording
Each call includes:
- Full transcription of the conversation.
- Automatic summary generated by the agent.
- Call recording.
This enables an advanced level of control and continuous improvement:
- Detect errors in COT execution.
- Validate whether the agent followed the conversational flow.
- Confirm whether the user actually interacted or if only the agent spoke.
- Adjust copy, tone, pauses, and timing to reduce robotic perception.