# VoIP - Call Transcription

The **transcription** is a step performed after the call. It allows the conversation to be turned into text, making **reading, analysis, and research** of the topics discussed during the service possible.

{% hint style="danger" %}

### Requirements for using Transcription

The transcription feature uses the artificial intelligence service from **OpenAI**.

For it to work correctly, it is necessary to:

* Have the **Artificial Intelligence module active** in the system
* Configure the **OpenAI API key** in **Integrations**

Without these settings, call transcription will not be performed.
{% endhint %}

{% content-ref url="/pages/67e643942ec3b7300bed00b9a783aee22e25cd7c" %}
[Integrations - Chat GPT](/en/topics/integration/integrations-artificial-intelligence-apis/integrations-chat-gpt.md)
{% endcontent-ref %}

***

Below are the steps to access the **call history** and view a transcription.

{% embed url="<https://youtu.be/YhxccoIAufE>" %}

***

### Transcription

In the **Transcription**, there is a space to view the **Call summary**, which condenses the main information discussed in a direct and objective way.

![img1](https://i.ibb.co/qLXP18vb/img1.png)

In the tabs above, it is possible to change the view to **Script** or **Conversation**, allowing you to see the full call transcription.

* **Script**: organizes the content in a structured way.
* **Conversation**: shows the dialogue exactly as it happened between the operator and the contact.

![img2](https://i.ibb.co/ccXjcTnW/img2.png)

If necessary, the transcription area also has a **prompt field** on the side, where it is possible to ask questions to the **Artificial Intelligence** about the call.

The AI can answer both about **explicit information from the conversation** and about **implicit details** that may have gone unnoticed.

![img3](https://i.ibb.co/CpRmdNQ6/img3.png)

***

## Scores

The scores analyze different aspects of call quality.

![img4](https://i.ibb.co/MxHJ9NhL/img4.png)

**Objectivity**\
Evaluates the consistency of the main topic discussed in the call.

**Resolution**\
Measures the quality of the solutions found for the problem presented.

**Engagement**\
Analyzes the level of interest and participation of the contact during the call.

**Assertiveness**\
Evaluates how direct and precise the operator was when addressing the topic.

**Clarity**\
Measures the ability to communicate clearly and understandably.

**Language tics**\
Analyzes the frequency of repeated words or expressions without informational value.

**Persuasion**\
Evaluates the operator's ability to convince the contact to make a decision.

***

## Metrics

![img5](https://i.ibb.co/hJNNGWDR/img5.png)

**Total duration**\
Total call time.

**Total words**\
Total number of words spoken during the conversation.

**Talk time**\
Time during which there was active speaking, excluding moments of silence.

**Operator / Contact (time)**\
How long each participant spoke during the call.

**Words per person**\
Number of words spoken by each side of the conversation.

**Speed (WPM)**\
Speaking pace measured in **words per minute**.

**Silence time**\
Pauses during the conversation, which may indicate **hesitation, reflection, or lack of engagement**.

***

## Charts

The charts help visualize call behavior.

**Talk vs Silence**\
Shows the balance between active conversation and pauses.

**Participation by time**\
Indicates who dominated the conversation: **operator or contact**.

**Words spoken**\
Shows the proportion of words used by each participant.

***

## 💡 Insights

![img6](https://i.ibb.co/whjDp9pc/img6.png)

**Improvements**\
Suggestions on what the operator can do better in future calls.

**Actions**\
Recommended next steps, such as **follow-up, CRM updates, or new interactions**.

**Suggestions**\
Practical guidance to improve service and **increase conversion chances**.

***


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.sprinthub.com/en/topics/general/voip/voip-call-transcription.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.