Use ChatGPT Work to build your data agent

# Use ChatGPT Work to Build Your Data Agent: A Setup Guide **TL;DR:** This guide explains how to set up ChatGPT as a capable data agent by connecting it to the same data sources and tools your team uses, primarily through the plugin system. ## Preparing the Agent's Context A useful approach is to think of the new AI agent as you would a new analyst joining your team. You should provide it with the same background information and resources you would give a human colleague. The first step is ensuring ChatGPT uses and connects to the same tools you use for your own work. ## Setting Up the Plugins To begin the configuration, navigate to the **Plugins** tab, located in the upper left of the interface. For this demonstration, a specific plugin called the **Data Agent** is used. Installing plugins is simple: click the plus button and log into your account if prompted. Users on Business or Enterprise plans may need to ask an administrator to enable certain plugins. ## Integrating Data Sources The data agent plugin has been configured to point to several different data sources: * A data table within **Databricks**. * An analysis channel within **Slack**. * Documents in **Google Drive** containing site metrics information. This setup allows the data agent to search for information across these sources and consolidate it into a single, coherent data context skill. ## Finalizing the Skill for Team Use The resulting data context skill is what appears on the right side of the interface. It is crucial that team members familiar with the data review and optimize this skill. Once verified as accurate, this skill can be packaged into a plugin for use across the entire company. Source: [https://www.youtube.com/watch?v=cXUmFec2PNY](https://www.youtube.com/watch?v=cXUmFec2PNY)