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

# Post Call Data Analysis

> Define custom data extraction from call transcripts

<img style={{ borderRadius: '0.5rem' }} src="https://mintcdn.com/subverse-611dde60/Shvr0mC998UZU1Eq/images/agents/post-call-analysis.png?fit=max&auto=format&n=Shvr0mC998UZU1Eq&q=85&s=63f5918d754f795746e695f6200822a1" width="964" height="1138" data-path="images/agents/post-call-analysis.png" />

Define custom data extraction from call transcripts.

### LLM Model

Select model for analysis:

* **OpenAI \[GPT 4o Mini]**: Fast, cost-effective
* **OpenAI \[GPT 4o]**: More accurate analysis

### Data Fields

**Type**
Data type to extract:

* Text
* Number
* Boolean
* Selector
* Array

**Name**
Field identifier (e.g., "summary", "customer\_sentiment", "order\_id")

**Instructions**
Tell the AI what to extract:

```
Summary of discussion between agent and customer. 
If no conversation found write N.A
```

**Examples:**

* **Lead Quality**: "Rate the lead quality from 1-10 based on interest level"
* **Next Action**: "What follow-up action is needed?"
* **Customer Sentiment**: "Was the customer satisfied, neutral, or dissatisfied?"
* **Order Details**: "Extract order ID if mentioned"

<Tip>
  Use post-call analysis to automatically populate your CRM, trigger workflows, or generate reports without manual review.
</Tip>