All Integrations
ChatGPT
x

Connect ChatGPT to Pigment and get governed answers

Connect ChatGPT to Pigment through the Pigment MCP Server, so answers about your Metrics, models, and plans come from governed Pigment data. Finance, Supply Chain, HR, and Sales teams can ask questions without uploading spreadsheets to an AI tool, and with advanced mode, enter data and build models. IT keeps every session within the access rights people already have in Pigment.

ChatGPT

ChatGPT

Connect ChatGPT to Pigment so teams can ask questions of governed planning data in natural language

Category

AI

Integration method

MCP Server

Best for

Planning analysis · Variance analysis · Model building

See documentation

What can you do with ChatGPT and Pigment?

  • Run variance analysis in plain language. An FP&A analyst asks why actuals are drifting from budget, and ChatGPT pulls the Metric data from Pigment without anyone building a report first.
  • Trace a number back to its drivers. Ask what feeds a Metric, and ChatGPT walks through its structure and dependencies in Pigment.
  • Find the model behind a forecast. Ask ChatGPT where a forecast lives, and it searches the Applications, Blocks, and Boards you can access.
  • Build and edit models from with natural language. In advanced mode, model builders have ChatGPT create Metrics, write formulas, or design Boards within their permissions.

How the ChatGPT integration works

To set it up, a Pigment Security Admin, Workspace Admin, or Primary Owner turns on the MCP Server to generate the Workspace's endpoint URL. A ChatGPT admin or owner adds that URL as a Pigment connector and publishes it, and each user signs in with their Pigment account.

The connection is read-only by default: ChatGPT can find Applications, Blocks, and Boards, read Metric structure, and retrieve Metric and List data. Each user can turn on advanced mode in Pigment to let ChatGPT enter data and build and edit models. Every request, read or write, runs under that user's Pigment permissions.

Key capabilities

  • Keep planning data in Pigment. ChatGPT reads directly from the models behind your forecasts and reports, so nobody uploads spreadsheets and there’s no second copy to reconcile.
  • Exclude sensitive Blocks. Turn off AI data access on any Block to keep it out of ChatGPT, while the rest stays available.
  • Block access to entire Applications. Remove the Use AI Agents permission from an Application, and ChatGPT can’t access it.
  • Roll out once for the whole team. A ChatGPT admin or owner publishes the connector to all or selected users, who skip Developer Mode.
  • Control which write actions run. ChatGPT can ask for confirmation before write actions, and admins on Enterprise and Edu plans can choose which actions the connector allows.

Frequently Asked Questions

What Pigment data can ChatGPT see?

ChatGPT sees only the Applications, Blocks, and data the signed-in user can already access in Pigment, and the connection grants no extra rights. Each Block must have AI data access turned on, and ChatGPT can read Metrics with Number, Integer, or Boolean data types.

Can ChatGPT write data or build models in Pigment?

Yes, in advanced mode, which is off by default. Each user turns it on in Pigment under Settings > Advanced features > Enable advanced tools on MCP server. ChatGPT can then create Applications, build Blocks, write formulas, design Views and Boards, and input data, but only within that user's permissions. For example, creating a Metric fails if the user doesn't have the Configure Blocks permission.

Will ChatGPT's answers match the numbers in our Pigment model?

ChatGPT retrieves Metric and List data from your live Pigment models, so answers draw on the same governed numbers and business logic your model uses. It can also read a Metric's structure and dependencies, which helps you check how a figure was built.

See Pigment in action

The fastest way to understand Pigment is to see it in action. Sign up today and explore how agentic AI can transform the way you plan.

Three colleagues focused on an iMac screen in a bright office with plants and modern artwork.

From 8 days to 4 min

Update P&L actuals & financial forecasting

80%

Time cut on data aggregation

12 hours

Saved per month on executive reporting

6 days faster

For scenarios creation and analysis