# Integrated Agent

In this section, we will demonstrate how to use the **Supervisor Node** within the **Multi-Agent** framework to flexibly orchestrate the two previously created agents—the **Web Search Agent** and the **Knowledge-Based QA Agent**. This enables a combined capability that leverages both internet-based search and internal knowledge retrieval to deliver more comprehensive answers.

## Overview of Flow Structure

<figure><img src="https://4046886348-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYf4L4Z4IuV1CQn3whQYb%2Fuploads%2FmKQDnirMlQlENHwH6jdJ%2Fcae36d308d316ba35f8454aa228881cb.png?alt=media&#x26;token=1cfb7499-fb33-47fc-a60d-7fbefb6f347d" alt=""><figcaption></figcaption></figure>

Under the default Multi-Agent configuration, you can directly utilize both agents by adding them to the **Supervisor Node**’s management scope. For detailed steps on how to add agents, please refer to the *Multi-Agent* section.

## Output Preview

<figure><img src="https://4046886348-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYf4L4Z4IuV1CQn3whQYb%2Fuploads%2FyU4ZGKGpD2vyxbhdmU3f%2F1189010562a10cbb17e622a2cf1679cd.png?alt=media&#x26;token=b646e540-f74d-4368-ad8c-3cf2806ec2d7" alt=""><figcaption></figcaption></figure>

<figure><img src="https://4046886348-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FYf4L4Z4IuV1CQn3whQYb%2Fuploads%2FPn7FYTS7ArCmTPqldgc2%2F4eab2fa90f656598340f8234b06f0c34.png?alt=media&#x26;token=55c53a51-8d0b-4bc5-8bbd-c53025113962" alt=""><figcaption></figcaption></figure>

Once the configuration is complete, users can input a natural language question, and the system will automatically route it to the appropriate agent:

* If the question is related to **WorkFX**, the **Knowledge QA Agent** will be triggered to generate an answer based on the internal knowledge base.
* If the question requires **general web search**, the **Web Search Agent** will be used to retrieve information from the internet.
* The final output will be presented in a unified format, ensuring users receive accurate and relevant answers.
