Configuring Workplace Assistant

Install and configure the Workplace Assistant in a Cloud Pak for Business Automation environment.

About this task

Use the following procedure to enable the AI assistant in a Cloud Pak for Business Automation environment.

For usage information, see Using Workplace Assistant.

Procedure

  1. During the standard installation of Cloud Pak for Business Automation, run the cp4a-prerequisites script in property mode, and select either 5a for Workflow Authoring, 5b for Workflow Runtime, or 8 for Workflow Process Service Authoring. Then, from the optional components list, select 4 to enable Workplace Assistant.
  2. In the generated property file, fill out the required fields as shown:
    WFA.WATSONX_API_KEY="<Required>"
    WFA.WATSONX_URL="<Required>"
    WFA.RUN_WORKPLACE_AGENT="<false>"
    WFA.WATSONX_PROJECT_ID="<Required>"
    IBM SaaS credentials, which appear by default:
    ###################################################
    ## USER Property for Workflow Assistant ##
    ###################################################
    ## Provide the API Key for WatsonX
    WFA.WATSONX_API_KEY="<Required>"
    
    ## Provide the Project ID of WatsonX
    WFA.WATSONX_PROJECT_ID="<Required>"
    
    ## Provide the URL of WatsonX
    WFA.WATSONX_URL="<Required>"
    
    ## Flag to run the Workplace Assistant
    WFA.RUN_WORKPLACE_AGENT="true"
    
    ## Flag to run the (Preview) Authoring Assistant
    WFA.RUN_AUTHORING_AGENT="false"
    
    ## Provide the CP4D user name.
    WFA.WATSONX_USERNAME="<Optional>"
    
    ## Provide the CP4D token.
    WFA.WATSONX_TOKEN="<Optional>"
    
    ## Provide the CP4D instance ID.
    WFA.WATSONX_INSTANCE_ID="<Optional>"
    
    ## Provide the CP4D version.
    WFA.WATSONX_VERSION="<Optional>"
    
    ## Optional - Only required if WATSONX_API_KEY is not set.
    WFA.WATSONX_PASSWORD="<Optional>"
    
    ## Optional - The ID of the LLM model to use. Default set to meta-llama/llama-3-3-70b-instruct.
    WFA.WATSONX_MODEL_ID="<Optional>"
    
    ## Optional - Only required if a custom deployment LLM model is being used. If set, will override WATSONX_MODEL_ID value.
    WFA.RUNTIME_AGENT_WATSONX_MODEL_ID ="<Optional>"
    WFA.AUTHORING_AGENT_WATSONX_MODEL_ID ="<Optional>"
    Example:
    WATSONX_URL=https://us-south.ml.cloud.ibm.com
    WATSONX_API_KEY=<your API key>
    WATSONX_PROJECT_ID=<your project id>
    RUN_WORKPLACE_AGENT=true
    
    For more information about property values, see Enabling generative AI in Business Automation Workflow on containers.
  3. Run the prerequisite script in generate mode and, when it completes, a Kubernetes secret named ibm-workflow-assistant-secrets is created, using the values that you provided.
    If you want to use LWE for the Workplace Assistant, patch the following secret:
    kubectl create secret generic ibm-workflow-assistant-secrets \
    
    --from-literal=WATSONX_URL="<your-watsonx-url>" \
    
    --from-literal=WATSONX_USERNAME="<your-watsonx-username>" \
    
    --from-literal=WATSONX_PASSWORD="<your-watsonx-password>" \
    
    --from-literal=WATSONX_VERSION="<your-watsonx-version>" \
    
    -n <namespace> \
    
    --dry-run=client -o yaml | kubectl apply -f -
    To go back using watsonx SaaS, use:
    kubectl create secret generic ibm-workflow-assistant-secrets \
    
    --from-literal=WATSONX_API_KEY="<your-api-key>" \
    
    --from-literal=WATSONX_URL="<your-watsonx-url>" \
    
    --from-literal=WATSONX_PROJECT_ID="<your-project-id>" \
    
    -n <namespace> \
    
    --dry-run=client -o yaml | kubectl apply -f -
  4. For the deployment that you chose, when the deployment script completes running, a lombardi-custom-xml-secret secret is generated, which contains the XML snippets that are required for the Workplace Assistant, and is referenced in the generated custom resource (CR) file.
    Note: If the secret lombardi-custom-xml-secret is already present, it is deleted and the old contents are stored as a backup XML file in the same location.
    Additionally, a new section is appended to the main CR file:
    workflow_assistant_configuration:
      run_authoring_agent: true
      run_workplace_agent: true
      
    In Workplace Assistant, the Workplace basic URL chat is automatically enabled.
    Important: In a federated environment, you must enable Workplace Assistant only on the Business Automation Workflow instance that hosts the federated Workplace.
  5. (Required only when you use Workplace in Navigator) If you want Workplace Navigator URL to have the web chat enabled, configure the parameters WorkplaceConfig.agent.enable and WorkplaceConfig.agent.endpoint as follows.
    1. Log in to IBM® Content Navigator and go to Administration in the navigation panel.
    2. Click Connections and, under Platform, select APPENG0.
      APPENG0 connection in the Navigator administrative console.
      Click Edit to enable the Connect button for APPENG0.
      APPENG0 connection with Connect button enabled.
      Click Connect and then go to the Applications tab, find the Workplace application and click Details to open the application.
      Workplace application in the Navigator admin console, under APPENG0.
    3. In the Workplace application, under Environment variables, set parameters as follows:
      The Environment variables tab of the Workplace application lists the WorkplaceConfig.agent.enable and WorkplaceConfig.agent.endpoint parameters.
      • Set WorkplaceConfig.agent.enable to true.
      • Set WorkplaceConfig.agent.endpoint to the agent endpoint. For example, https://<cpd_url>/agent/runtimeChat.
        Important: The URL is case-sensitive. Ensure that you use the correct capitalization in the final path (/runtimeChat). For example, https://<cpd_url>/.../runtimeChat.
      Also, make sure you use the namespace of your Cloud Pak for Business Automation deployment in the URL, and replace the OCP name with the cluster name.
  6. Optional: To disable the Workplace chat, in the main CR file mentioned in step 4 earlier, set run_workplace_agent to false.
  7. Optional: To enable Content Assistant features for content search, querying, and summarization, see the following configuration topic for setup instructions and prerequisites.

Results

After you configured Workplace Assistant, you can start using it in your work. See Using Workplace Assistant.