Getting Started with Prophecy DataGPT

Prophecy DataGPT lets you work with your data by chatting, right inside ChatGPT, Claude, Copilot, or Prophecy. This guide takes you from installation to starting your first analysis.

1. Set up DataGPT

Option 1: Start from your AI assistant

  1. Log in to your account on ChatGPT, Claude, or Copilot

  2. Install the Prophecy DataGPT connector by following the documentation here

  3. After connecting, you will be prompted to sign up or sign in to Prophecy. Once complete, you are set up to use DataGPT in your AI Assistant

Option 2: Start from Prophecy

  1. Sign up or sign in at app.prophecy.ai

  2. Navigate to the Chats page. This page has the same DataGPT functionality as the connector.

2. Add a file

Upload a file by dragging it into the chat. Prophecy supports XLSX, CSV, and other formats. Then you can prompt:

Add this file to my Databox

If you need a sample dataset, use this one:

employee_attrition_sample_data.csv (699.1 KB)

Your AI assistant will automatically upload the file into your Databox, which is a private space where Prophecy stores files and connections. If you want to add direct data connections to sources like SharePoint, AWS, etc., you can set that up here.

3. Start an analysis with DataGPT

Now you’re ready to start analyzing! You can try prompts like:

Explore the data and summarize important statistics

Build a workflow to clean this dataset

Create charts to visualize my data

Note: Upon your first use of DataGPT, you will be prompted by your AI assistant to allow the DataGPT connector to take actions. Click “always allow” if you would not like to receive this prompt each time

4. Review your workflows and analysis

For analysis visuals, click “Audit” to see how values were calculated:

This will send a prompt to pull up the Audit view, where you can see the data in and out at each step in the workflow. Click on “See Trace” to pull up the step-by-step view:

If you want to make edits, simply prompt with your edit request in plain language. For example: add employee grouping by gender

To learn more about what DataGPT can do, check out all the features in our documentation.

Start a community thread or respond below if you have any questions!