AI Call Analysis
After every call, Akol’s AI analyzes the conversation and generates a detailed summary. This helps you understand what happened without listening to the full recording.
Where to Find It
- Go to Dashboard > Calls
- Click on any completed call
- The AI analysis appears in the call detail page
Call Summary
Every call gets a natural language summary — a 1–2 paragraph overview of what happened during the conversation. It covers:
- What the caller wanted
- What information was provided
- What actions were taken (appointments booked, messages sent, etc.)
- How the call ended
Outcome Classification
Each call is classified into one of six outcomes:
| Outcome | Description |
|---|---|
| Successful | The call achieved its goal — appointment booked, question answered, issue resolved |
| Partial | Some goals were met but not all (e.g., question answered but caller didn’t book) |
| Callback Needed | The caller requested a callback or needs follow-up from your team |
| Failed | The call didn’t achieve its goal — caller was dissatisfied or the AI couldn’t help |
| Spam | The call was identified as spam or an unwanted call |
| Dropped | The call was disconnected unexpectedly |
Customer Intent
The AI identifies the caller’s primary reason for calling. Examples:
- “Schedule a dental cleaning”
- “Ask about pricing for a consultation”
- “Reschedule an existing appointment”
- “File a complaint about service”
Sentiment Analysis
Each call receives a sentiment classification:
| Sentiment | Description |
|---|---|
| Positive | Caller was satisfied, friendly, or enthusiastic |
| Neutral | Caller was matter-of-fact, neither positive nor negative |
| Negative | Caller was frustrated, upset, or dissatisfied |
The sentiment also includes a score (a numeric value) for more granular tracking.
Emotion Tags
Beyond overall sentiment, specific emotions detected during the call are tagged. Examples:
- Satisfied, grateful, enthusiastic (positive)
- Confused, uncertain (neutral)
- Frustrated, angry, impatient (negative)
Actions Taken
The analysis lists every action your AI agent performed during the call:
- Booked an appointment
- Sent an SMS confirmation
- Looked up customer information
- Transferred the call
- Took a message
Follow-Up Recommendations
If the AI determines follow-up is needed, you’ll see:
- Follow-up needed flag
- Reason — Why follow-up is recommended
- Suggested deadline — When the follow-up should happen
Use the follow-up recommendations to prioritize your callback list. Calls marked as “Callback Needed” with a follow-up deadline should be handled first.
Using AI Analysis
Prioritize Your Day
Sort calls by outcome to quickly find:
- Callback Needed calls that require your attention
- Failed calls to understand what went wrong
- Negative sentiment calls that might need damage control
Improve Your Agent
Look at patterns in your call analysis:
- Are certain questions consistently resulting in Partial outcomes? Update your agent’s knowledge
- Are callers frequently Frustrated? Adjust your agent’s tone or responses
- Are many calls being Dropped? Check your connection quality
Track Trends
Over time, the AI analysis helps you spot trends:
- Is customer satisfaction improving or declining?
- Which types of calls have the highest success rate?
- What are the most common reasons people call?