Workday Planning MCP Server

One governed model.
Every orchestrator.

The Workday Planning MCP server exposes the same governed planning model to Claude, Gemini, Cortex, and any MCP-compliant client. Your governance — your security, your permissions, your audit trail — travels with every request, so AI tools act on the live plan without ever stepping outside it.

Featured reference implementation

A partner-built standalone Claude app that generates specialized data-entry forms, writing straight back into the governed plan through the Workday Planning MCP server.

From the team

“Same governance, new surface area. The MCP server lets any AI tool reason over the live planning model — under the same identity, permissions, and audit trail as the product itself.”

Workday Adaptive Planning

MCP Server team

How it works

A planning team
wiring an MCP client into their stack.

Same governance, new surface area

The Workday Planning MCP server exposes the governed model to any orchestrator your team chooses — no parallel system, no shadow data set.

A planner asks Claude a question about Q4 headcount; the answer is computed against the same governed model as the Adaptive Planning UI, with the same security applied.

Works with Claude, Gemini, and beyond

Compatible with any MCP-compliant client — no bespoke adapters, no custom plumbing. Drop the server into the orchestrator your team has already chosen.

The same MCP server powers a Claude desktop app for finance, a Gemini agent in workflow automation, and a custom Cortex client for analysts.

Identity, scopes, and audit built in

Every tool call carries the caller's identity and is logged against the same audit trail as the product. There is no anonymous access path.

A regional FP&A lead's MCP calls are scoped to their region, and every read or write shows up in the audit trail the same way a UI action would.

Compatible clients

What teams build
on the MCP server.

Use Case 01

Claude desktop app for FP&A

For FP&A teams using Anthropic's MCP client and Claude Skills

A finance team uses Claude as their day-to-day reasoning surface. With the Workday Planning MCP server connected, Claude can inspect sheets, explain variances, and draft reforecasts against the live governed model — under each analyst's own permissions.

Analysts stay in their preferred AI tool without leaving the governed plan
Every Claude tool call is identified, scoped, and audited
Skills can call MCP tools to read accounts, dimensions, and versions live

Use Case 02

Gemini agents in workflow automation

For platform teams building Vertex agents on Google Cloud

Gemini agents orchestrate multi-step planning workflows — pulling pipeline, calling the MCP server for the latest plan, and writing back proposed updates. The same governance applies whether the agent runs interactively or on a schedule.

Agents read and write the governed plan without bespoke connectors
Service accounts use tenant-scoped tokens with row-level security
Scheduled runs land in the same audit log as a UI user

Use Case 03

Snowflake Cortex reasoning over the plan

For analytics teams running Cortex alongside Adaptive Planning

Cortex blends warehouse data with live planning context via the MCP server, so analyst-facing assistants can compare actuals, scenarios, and pipeline without exporting the plan to the warehouse.

Live planning context, not a stale extract, joins warehouse analysis
No new ETL pipeline to maintain between the plan and the warehouse
Read paths respect the user's Adaptive Planning permissions

Use Case 04

Custom in-house MCP client

For platform engineers building bespoke planning experiences

Teams build their own MCP-compliant clients — embedded copilots, specialized data-entry apps, ops bots — and connect them to the same governed model. No custom Workday SDK to ship, no parallel API to keep in sync.