The Workiva MCP connects Claude to your Workiva content so it can read and analyze your documents and data on your behalf. You ask a question in Claude, and it reads across your Workiva documents and data, runs the analysis, and responds with the sources it used.
This guide walks you through connecting the Workiva MCP to Claude, approving access, and asking your first question grounded in your Workiva content.
Workiva MCP is a custom connector and won't appear when searching in Claude by default. Your Claude administrator needs to add it as a custom connector before you can access it. Once added, it appears alongside your other custom connectors.
This feature is currently available through the Early Access Program (EAP) until September 14th, 2026. The preview feature is provided to you for technical testing and evaluation. Workiva may extend the testing period of the preview features at its discretion. If your team is interested in using this feature, please contact your Customer Success Manager.
Prerequisites
Before you begin, ensure you have the following:
- An active Workiva account with access to the workspaces you want to interact with through MCP.
- Access to claude.ai (Claude is a supported connect path at EAP launch).
- Your organization's security admin notified that they will need to grant your AI agent access to specific workspaces via the Connected AI agents console after you connect.
At launch, the Workiva MCP is read-only. Claude can read documents and retrieve data from Workiva, but cannot create, edit, update, or delete artifacts.
Step 1: Add the Workiva MCP connector in Claude
- Open Claude Desktop, navigate to Add > Connectors > Browse connectors.
-
Search for Workiva MCP and click on it.
-
Select Connect.
-
You'll be redirected to the standard Workiva sign-in page. Use the same credentials you use to log in to Workiva.
-
A confirmation screen will appear showing what Claude will be able to access on your behalf. Review and approve.
- Once approved, Claude connects under your identity and acts on your behalf, only within your existing Workiva permissions. Claude can never access more than you can.
You don't need to select organizations, workspaces, or permissions at this step. Your organization's security admin controls workspace-level access separately via the Connected AI agents console.
Step 2: Admin grants workspace access
After a user connects, a security admin must grant the AI agent access to specific workspaces before the user can query their data.
- In Workiva, navigate to Connected AI agents in admin settings.
- Locate the agent connected by the user (for example, Claude).
- Grant or deny access per workspace as required.
- When access isn't granted, Claude will inform the user that it can't act in that workspace.
If a user connects but can't see their data, this is the most likely cause. See the following Troubleshooting section.
Step 3: Ask your first question
Before diving into real work, confirm Claude can see your Workiva data. Try this:
"Using Workiva's MCP show me the workspaces I have access to."
Once Claude returns your workspaces, follow up with:
"List the documents in [workspace name]."
If Claude returns results, you're connected and ready to go. If not, revisit your permissions setup before continuing.
| Use case | What you can ask |
|---|---|
| Earnings Q&A prep | Pull the current 10-Q, 10-K, earnings release, MD&A, and prior scripts, and surface a Q&A view with citations and consistency flags. |
Troubleshooting
- I connected but can't see my data. This is almost always an access issue. Your organization's admin needs to grant your AI agent access to specific workspaces in the Connected AI agents console. Ask your admin to grant access for the workspaces you need.
- Claude says it can't act in a workspace. An admin has restricted the agent for that workspace. Contact your admin if you need access.
- My session expired. Access tokens are short-lived by design. Reconnect and sign in again to start a new session.