![[Runyour Guide #2] Building Your Own AI Work Agent Using Internal Documents](https://cdn.prod.website-files.com/659faaa08c491caeb4e3f8a3/6a91304397f7625a48a82183_%EB%A9%94%EC%9D%B8_260824_%ED%8A%9C%ED%86%A0%EB%A6%AC%EC%96%BC%202.png)
Hi, this is the Runyour team.
Do you find yourself repeating the same instructions every time you assign a task to AI?
"Our company sells these products, and our press releases follow this format. Please write this based on those materials."
If the AI lacks context about your company and workflows, you have to re-explain the necessary information every single time.
With Runyour Agent’s knowledge base and group agent features, you can connect your internal documents for the AI to reference and create specialized agents for specific tasks to use repeatedly.

Watch the full process in the video, from uploading documents and creating a knowledge base to setting up your agent and generating actual content.

First, go to [Group] → [Knowledge Base] to upload the materials you want the AI to reference.
For example, if you are building an agent to write press releases, you can use the following materials:
The materials you register this way can be used by the AI to search for and reference the information it needs to generate answers.
In other words, rather than retraining the model itself, it is more about making your organization's data accessible and usable whenever you need it.
Once your knowledge base is ready, you can create an agent to handle actual tasks.
You can define the agent's role and how it should provide responses.
For example, if your goal is to write press releases,
and more, then connect the knowledge base you created earlier to the agent.
You can follow this process step-by-step through the actual setup screens in the video.
Once setup is complete, you can request tasks immediately without having to explain your company or work context from scratch every time.
For instance, if you provide information about a new product launch, the agent will draft a press release by referencing the connected materials and the defined writing style.
Once the output is ready, you can simply review and edit the necessary parts to use it for your actual work.

If you want to get more out of Runyour, check out the following tutorials.
📍 [Part 1] Getting Started with Runyour Agent: From Sign-up to Dashboard
📍 [Part 3] Integrating Runyour API with VS Code for AI-Powered Coding