Installing Workflow Runtime MCP Server

Edit online

Workflow Runtime MCP Server is installed as an optional component during the Business Automation Workflow installation process.

Before you begin

  • Complete the cluster administrator setup by running ./baw-clusteradmin-setup.sh.
  • Ensure that you have a namespace created for your Business Automation Workflow deployment.

About this task

Workflow Runtime MCP Server enables AI-powered automation for your workflow applications through standardized integration with Large Language Models and AI agents. When you select Workflow Runtime MCP Server during installation, the generated Custom Resource (CR) includes the configuration for Workflow Runtime MCP Server. Upon applying the CR, all Workflow Runtime MCP Server-related operands are deployed, and the relevant pods start.

The server supports both Process and Case management capabilities, enabling AI agents to:

  • Search for processes and tasks based on specific criteria
  • Claim, update, and complete tasks
  • Manage process instances and case workflows
  • Access and update business data
  • Interact with documents associated with processes and cases
Note: Workflow Runtime MCP Server supports deployments with and without IBM® Process Federation Server. You do not need to install Process Federation Server or Federated Data Repository (FDR) to use Workflow Runtime MCP Server.
  • If Workflow Runtime MCP Server is configured to use Process Federation Server, all queries route through Process Federation Server.
  • If Process Federation Server federates only some instances, the remaining instances use their respective MCP servers.
  • Tools that affect specific instances route directly to the respective runtime.
  • Traditional environments that Process Federation Server federates are not supported in this release.

Procedure

  1. Run the prerequisites script to continue the installation process. 
    ./baw-prerequisites.sh -m property -n namespace

    Replace namespace with the name of your Kubernetes namespace.

  2. When prompted, select the optional components to include in your deployment.

    The script displays a list of available optional components. To install Workflow Runtime MCP Server, select Workflow Runtime MCP Server from the list.

  3. Complete the remaining prompts in the prerequisites script to generate the CR file.

    The script generates a Custom Resource (CR) file that includes the Workflow Runtime MCP Server configuration.

  4. Apply the CR to deploy Business Automation Workflow with Workflow Runtime MCP Server. 
    kubectl apply -f ibm_cp4a_cr_final.yaml
  5. Verify that the Workflow Runtime MCP Server pods are running. 
    kubectl get pods -n namespace | grep mcp

    Wait until all Workflow Runtime MCP Server pods show a status of Running.

  6. Retrieve the Workflow Runtime MCP Server access information from the ConfigMap. 
    kubectl get configmap bawdeploy-cp4ba-access-info -n namespace -o yaml

    After Workflow Runtime MCP Server is enabled and all relevant pods are running, the system updates the bawdeploy-cp4ba-access-info ConfigMap with access details.

What to do next

For information about configuring the Workflow Runtime MCP Server, see Managing work with the IBM Workflow Runtime MCP Server.

For information about customizing the Workflow Runtime MCP Server configuration parameters, see: