About the Adaptive Planning MCP Server
To request this feature, contact your Named Support Contact to submit a Workday Customer Care request. After confirming that you're eligible, we'll enable the feature.
The Adaptive Planning Model Context Protocol (MCP) Server provides a secure, standardized interface for external customer and partner AI agents to read Adaptive Planning data and model context. MCP enables agents to discover available tools, request permission-aware planning data, and combine Adaptive Planning context with external operational or market data without custom point-to-point integrations.
The MCP Server supports read-only customer and partner use cases such as:
- Exploring sheet data by account, time period, version, level, and dimension.
- Retrieving metadata for accounts, levels, versions, time periods, sheets, and custom dimensions.
- Comparing financial data across versions, including budget, forecast, actuals, and scenarios when the user has access.
- Generating narrative reporting, charts, and variance insights in an external AI workspace.
- Exporting raw data or dimensions when the user has the required permission set. Identifying rows or planning records that meet conditions, such as expenses over a threshold or employees that match workforce planning criteria.
- Enabling partner agents to combine Adaptive Planning data with external data sources for unified reporting, variance analysis, and driver-based planning insights.
Data Flow
At a high level, requests follow this flow:
- An external customer or partner AI agent sends a MCP request.
- The AI agent calls the MCP server to get a list of tools. The MCP server returns the tools list based on user permissions.
- The AI agent determines which set of tools to call to respond to the user query. It then calls the tools.
- Once the MCP tool returns the data, the AI agent parses and presents the data in the requested format.
MCP Server Components
The MCP server components include:
- Host (customer) : The external AI application or agent that receives the user’s request and orchestrates the interaction. Examples: Claude, VS Code, Gemini CLI, or Cursor IDE.
- Client (customer): The component inside the host that communicates with the MCP Server by using the MCP protocol.
- Server: The Adaptive Planning MCP Server that exposes Adaptive Planning capabilities to the client.
- Tools: Callable functions that the agent can use, such as reading sheet data, retrieving metadata, or comparing versions.
- Transport: The communication layer. The source docs mention JSON-RPC 2.0 and Streamable HTTP.
- Security and Governance: Authentication, authorization, input validation, output sanitization, rate limiting, audit logging, tool availability, and permission enforcement.
Agent Security
Agent Security helps AI agents operate securely within your enterprise by providing unique identities, robust authentication, and granular authorization controls. Agent Security governs how agents access data and perform tasks according to the principle of least privilege.
Workday secures AI agents using the same robust framework used for users. The product creates a trusted environment through two layers of protection:
- Governance: Treats agents as nonhuman identities, helping prevent agent sprawl and ensuring that every machine action is associated with a specific, manageable account.
- Transparency: Every action an agent performs is fully auditable, providing the same visibility into agent activity that you have into employee activity.
You manage agent definitions and lifecycles in the Agent System of Record (ASOR). Agent Security governs agent behavior at runtime.
The Agent Security framework currently supports Delegate execution mode. The agent acts on behalf of a user. Example: A user delegates to an agent the task of requesting time off for a specific date.
Agent Identity: Agent System User (ASU)
Agents use Agent System User accounts ((ASUs). Each agent has a unique identity, ensuring unambiguous control over its access and providing transparency for auditing.:
- Unique identity: Each agent has a unique ASU account.
- Management: Adaptive Planning automatically creates and manages ASUs when you configure skills in the ASOR. Example: Activating an ambient skill triggers the creation of the required ambient ASU.
ASU Creation and Registration
ASU creation occurs as part of agent configuration and skill enablement in ASOR:
- ASU generation: When an administrator enables skills for an agent, ASOR automatically creates a dedicated Delegated ASU or Ambient ASU for that agent in a disabled state. When the agent is activated, Workday automatically enables the associated ASUs.
- OAuth client and key pair: Adaptive Planning generates an OAuth client and associates it with an ASU. The product also creates a private key pair and registers the public key with the Workday authorization server. The ASU can then authenticate and perform the tasks defined by its skills.
- Credential storage: ASOR securely stores the client credentials, including the private key, in the Credential Store. ASOR retains a unique reference ID that points to the key-pair credentials. These credentials generate ASU tokens. For a Delegated ASU, Adaptive Planning exchanges the ASU token together with the user token rather than through OAuth 2.0 client credentials.
MCP Configuration and Connection
To configure external agents for standalone Adaptive Planning, see:
To configure external agents for Workday-enabled Adaptive Planning, see: