# ChatGPT - Allow the model to take control of the conversation to make it easier to interact with the lead

<figure><img src="/files/f65fca9dfffdd94f83fdd3faa937b4f72c3354c6" alt=""><figcaption></figcaption></figure>

### Introduction

Sprinthub allows the **GPT** to take control of the support service to facilitate interaction with customers. This feature can be configured to allow the model to start and resolve support sessions automatically based on the available information. The GPT's control of the service can be adjusted according to business needs, ensuring efficiency and personalization.

### How to Enable GPT Service Control

To activate this feature, simply check the option **"Allow to control the service"** within the model settings. When enabled:

<figure><img src="/files/4b69a6975f99cd1f816e3d9aaa597ad7b4cf9d71" alt=""><figcaption></figcaption></figure>

* The GPT can start conversations automatically.
* It can answer questions and provide information based on the available context.
* It can close support sessions when the user's questions are resolved.

> **Note:** This feature is only available when the model is started from a chatbot within Sprinthub.

### Examples of Contextualization for the GPT

Contextualization is essential to ensure the GPT responds correctly to users. In Sprinthub, the level of contextualization can be adjusted according to the complexity of the support. Below are some examples of how to structure contextualization:

#### **1. Starting the Service**

The GPT can be configured to start a conversation with customers as soon as it detects an interaction or based on specific rules.

**Instruction example:**

> "Greet the user in a friendly manner and ask how you can help them. If the customer's question is related to products, provide details based on the available information."

**Expected response:**

> "Hello! Welcome to our support. How can I help you today?"

#### **2. Ending the Service**

To ensure the service is concluded correctly, the GPT can be instructed to close the conversation when it perceives that the user's need has been met.

**Instruction example:**

> "If the customer indicates that their question has been resolved or has no further questions, close the service politely and offer additional support if necessary."

**Expected response:**

> "I'm glad I could help! If you have more questions in the future, I'll be here. Have a great day!"

#### **3. E-commerce Support**

**Model configuration:**

* Goal: Assist customers in searching for products and purchase information.
* Restrictions: Do not answer questions that are not in the catalog.
* Instruction example:

  > "Act as a sales assistant. If a customer asks about prices or product specifications, use the registered information. If the question is not in the database, respond with 'Humm... I'm not sure.'"

#### **4. Technical Support**

**Model configuration:**

* Goal: Help users with technical questions about a software.
* Restrictions: Respond only based on the provided documents.
* Instruction example:

  > "Answer only based on the available technical manual. If the question is not covered, recommend that the user contact specialized support."

#### **5. Meeting Scheduling Assistant**

**Model configuration:**

* Goal: Schedule meetings based on the user's availability.
* Restrictions: Do not make decisions without the customer's confirmation.
* Instruction example:

  > "Assist the user in scheduling meetings by checking available times. Always ask if the suggested time is suitable before confirming."

### Conclusion

GPT service control in Sprinthub enables efficient and personalized automation. By adjusting contextualization correctly, it is possible to create intelligent assistants that improve the customer experience, ensuring precise and relevant interactions. Evaluate your needs and configure your model to optimize support in Sprinthub!


---

# 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/artificial-intelligence/ai-agents/legacy-ai-agents-notice/strategies/chatgpt-allow-the-model-to-take-control-of-the-conversation-to-facilitate-interaction-with-the-lead.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.