How It Works
Understanding Evalua's automatic evaluation process
Evalua operates automatically, analyzing calls without manual intervention (unless automatic call evaluation is turned off).
The Process
1. Call Recording
Your existing phone system records calls automatically.
2. Transcription
After the call ends, Evalua converts audio to text, labeling each speaker (agent and customer).
3. Analysis
The system reviews the transcription against your knowledge documents — including any Context Documents you've provided.
4. Evaluation
Each call receives a status. The most common ones:
- Passed — met all or most criteria
- Failed — did not meet quality standards
- To Review — the AI was uncertain; a supervisor needs to make the call
After a supervisor reviews a "To Review" call, it becomes Review Pass or Review Fail.
A handful of statuses indicate the call couldn't be evaluated end-to-end — for example Missing Recording, No Knowledge (the workspace has no knowledge documents yet), or Limit Exceeded (the company/user is over their transcription quota). See Evaluations for the full list.
5. Results
You receive a report with the full transcription, analysis of each criterion, specific feedback, and an overall summary.
Timeline
The entire process happens within 2-5 minutes of the call ending depending on call duration. Results appear in your Dashboard and History page automatically.
Accuracy
Evalua evaluates based on your specific knowledge documents — Overview, Feedback, and Context — not generic templates. As you refine them, evaluations become more precise. Calls the AI is unsure about are routed to To Review rather than guessed at, so supervisors keep the final say.