> ## Documentation Index
> Fetch the complete documentation index at: https://docs.movoice.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Observability & Analytics

> How to extract insights, sentiment, and structured data from your Voice AI calls.

# Observability & Analytics

Understanding what happens during a call is just as important as the call itself. Movoice AI provides deep, automated observability tools to help you measure success and improve your agents.

## Post-Call Analysis

Every call handled by Movoice AI undergoes automatic AI analysis within seconds of completion.

### 1. Automated Summaries

Instead of listening to hours of recordings, your team can read a concise, paragraph-style summary of the entire conversation.

* **Key Actions**: AI identifies if a booking was made, a lead was qualified, or an issue was escalated.
* **Sentiment Tracking**: Measures the caller's mood (Positive, Neutral, Negative).

### 2. Structured Data Extraction

Movoice AI can extract specific data points from conversations and sync them directly to your CRM or database.

* **Entities**: Names, phone numbers, email addresses.
* **Outcome Mapping**: Custom tags like `BOOKED`, `REFUSED`, or `WANTS_CALLBACK`.

## Real-Time Monitoring

For high-volume enterprises, real-time observability is critical for monitoring agent performance.

### Live Transcripts

View a live, scrolling transcript of any active call directly from the **Call Logs** dashboard. This allows for:

* **Instant QA**: Monitor how agents are handling complex queries.
* **Human In-The-Loop**: Identify when an agent might need a manual transfer.

### Performance Metrics

Track high-level KPIs across your entire fleet of agents:

* **Call Completion Rate**: % of calls that reached a definitive outcome.
* **Average Interaction Time (AIT)**: Average duration of successful calls.
* **Response Latency**: Real-world measurement of turn-around time.

***

## How to use Call Insights

You can access these insights via:

1. **The Dashboard**: Visual charts and searchable logs.
2. **Webhooks**: Receive a JSON payload as soon as a call ends.
3. **API**: Query the `/calls` endpoint to pull structured data into your own reports.

> \[!TIP]
> Use **Sentiment Analysis** to filter for "Negative" calls and proactively have a human agent follow up to save the relationship.