2026 Complete Guide to Setting Up Your Workday MCP Server

AI agents can now retrieve employee records, run payroll queries, surface org chart data, and check time-off balances — directly from Workday, in real time. The model context protocol (MCP) makes this possible by giving AI clients governed, schema-driven access to enterprise systems without replicating data or building custom APIs.

However, building an MCP server from scratch requires more than most IT teams want to take on: selecting and configuring an MCP SDK, writing tool schemas for each Workday object, implementing OAuth 2.0, hardening the endpoint for production, and maintaining all of it as Workday APIs evolve.

CData Connect AI removes that overhead. It provides a pre-built Workday connector and a production-ready MCP endpoint you can wire to any AI client in minutes.

Why a managed MCP platform is the right choice for enterprise

The DIY path to Workday-AI integration works, but it carries a long maintenance tail. A managed MCP platform like CData Connect AI abstracts the server layer entirely — authentication, schema management, compliance, and versioning are handled on your behalf.

The choice comes down to build versus connect. Here’s what each path involves:

For most enterprises, the managed path is faster to deploy, easier to govern, and less expensive to maintain.

Setting up Workday with CData Connect AI

The steps below walk through the full setup: from creating a Workday connection in Connect AI to verifying live data before you bring in any AI client.

Step 1: Create a Workday connection

1. Log into CData Connect AI and go to Sources.

2. Click Add Connection and select Workday from the source list.

3. Enter your Workday Tenant and BaseURL. To find these, log into Workday and search for “View API Clients.” The Workday REST API endpoint appears in the format https://domain.com/subdirectories/mycompany — the subdirectory portion is the BaseURL, and the final path segment is the Tenant.

4. Select your ConnectionType to set which Workday API to use: WQL, REST, SOAP, or Reports-as-a-Service (RaaS). WQL and RaaS cover most read-heavy HR and reporting use cases; SOAP gives more granular access to base data objects.

5. Save and test the connection.

Step 2: Select the Workday modules to expose

Choose which Workday objects your AI clients can access: Workers, Payroll, Reports, Absence, Time Tracking, or any other available objects. Start with read-only modules before enabling write operations.

Step 3: Test before connecting any AI client

Run a quick query in Connect AI’s built-in Data Copilot to confirm your Workday data is returning correctly before wiring in any external AI client.

1. In Connect AI, navigate to Data Copilot from the left menu.

2. Enter a natural-language question — e.g., “Show me all active workers in Workday” — and click Submit.

3. Expand the tool results (getCatalogs, getSchemas, getTables) to verify data is returning as expected.

Connecting your AI client to Workday

Once your MCP endpoint is live, connecting it to an AI platform takes a few steps. All four clients below use the same endpoint URL generated by Connect AI; authentication passes through your Workday credentials automatically, so no separate permission mapping is required.

Claude

1. Sign in to Claude.ai or create a new account.

2. Expand the left navigation menu.

3. In the bottom-left corner, click your username and select Settings.

4. In Settings, click Connectors, then Browse Connectors.

5. Search for CData Connect AI and click the listing.

6. Click Connect and sign into your Connect AI account.

7. After connecting, Claude opens a new chat — you can now query your Workday data.

Note: Claude’s setup process is subject to change. Refer to Claude’s documentation for the latest setup information.

Full walkthrough: Connect Claude to Workday.

ChatGPT

Note: MCP support in ChatGPT requires a Plus or Pro subscription.

1. Sign in to ChatGPT with a Plus or Pro account.

2. Navigate to Apps from the left panel.

3. Search and select CData Connect AI from the app list.

4. Click Connect to authenticate your Connect AI account.

5. Click Sign in with CData Connect AI to authorize.

6. After authorization, click Start Chat to begin a session with Connect AI running in the background.

Full walkthrough: Connect ChatGPT to Workday.

Microsoft Copilot Studio

1. In your Copilot Studio agent, go to Tools → Add tool → New Tool.

2. Search for and select CData Connect AI.

3. Click Create to authenticate via OAuth.

4. Click Add and configure to add the tool to your agent.

Full walkthrough: Connect Copilot to Workday.

Google Gemini

Note: The Gemini integration uses Gemini CLI. Ensure it is installed (npm install -g @google-gemini/cli).

Configure the Gemini CLI

1. Locate or create your Gemini CLI settings file (~/.gemini/settings.json on Mac/Linux, %USERPROFILE%\.gemini\settings.json on Windows).

2. Add the Connect AI MCP server entry, replacing YOUR_EMAIL and YOUR_PAT:

{
  "mcpServers": {
    "cdata-connect-ai": {
      "httpUrl": "https://mcp.cloud.cdata.com/mcp",
      "headers": {
        "Authorization": "Basic YOUR_EMAIL:YOUR_PAT"
      }
    }
  }
}

3. Save the file. Gemini CLI will now route data queries through Connect AI.

Full walkthrough: Connect Gemini to Workday.

Security, governance, and compliance

Connect AI passes Workday’s native access controls directly through to the AI layer — the AI agent acts as the authenticated Workday user and can only access what that user’s role permits.

Authentication runs on OAuth 2.0 with SSO pass-through, with no separate credential store to manage. Workday’s native role-based access control (RBAC) applies to every query — Connect AI does not bypass or re-map permissions. Every AI query is logged with user identity, timestamp, Workday objects accessed, and filters applied. CData Connect AI holds SOC 2 Type II and ISO 27001 certifications, covering the requirements for enterprise HR data.

Before going to production, enable audit logging and review it before onboarding any AI user, set query rate limits per user to prevent runaway API calls, start with read-only Workday modules and enable write operations only after a controlled pilot, and review Connect AI activity logs monthly.

Best practices for reliable Workday-AI integration

A few configuration choices make the difference between a reliable Workday-AI integration and a risky one:

• Start with read-only modules (Workers, Reports) before enabling write operations (Timesheets, Payroll). Fewer blast-radius risks during early deployment.

• Scope queries to the relevant team or cost center. Don’t expose all Workday data to every AI user — use Connect AI’s workspace and tool scoping to enforce boundaries.

• Require explicit confirmation before any write operation executes. Deletes and updates should never run without a human-in-the-loop approval step.

• Connect Workday alongside other enterprise systems in a single AI session. Common combinations: Workday and ServiceNow for cross-system HR and IT workflows, or Workday and NetSuite for HR and finance data together.

• Review Connect AI audit logs regularly to monitor AI activity on sensitive HR data.

Frequently asked questions

Connect your Workday data to AI today with CData Connect AI

Workday holds some of the most sensitive and operationally critical data in your organization. With CData Connect AI, your AI agents can access it in real time — governed, audited, and scoped to exactly what each user is permitted to see — without a single line of custom server code.

Start a free trial and connect your Workday data to AI today.