Configuring the HelixGPT Insight Finder agent


Configure the HelixGPT Insight Finder agent to use the AI capabilities in BMC Helix Dashboards for generating insightful and customized dashboards. The following configurations are required:
 

Perform these configuration steps in sequence to use HelixGPT Insight Finder.

 ConfigurationDetails
1Configure an AI model.

A model identifies the Large Language Model (LLM) and its version used in HelixGPT Insight Finder. HelixGPT Insight Finder supports specific models. Configure the required model by providing the API keys for authentication and specifying values for other configuration parameters.

For information about the models supported by HelixGPT Insight Finder, see Models in BMC Helix GPT.

2

Create sub-agents for the products that you are using, such as BMC Helix ITSM and BMC Helix AIOps.

3Create an Insight Finder Supervisor Agent.Create an Insight Finder Supervisor Agent by copying the configuration details from the existing out-of-the-box Insight Finder Supervisor Agent.
4Create a skill.Create a skill and add the Insight Finder Supervisor Agent to the skill. Copy the configuration details from the existing out-of-the-box Insight Finder Supervisor Skill.
5Configure the skill ID.Add the skill ID of the Insight Finder Supervisor skill to the HelixGPT Insight Finder settings in BMC Helix Dashboards. The skill is required to enable a conversational user experience for end users.
6Select specific views.Specify which custom or out-of-the-box views you want to use in HelixGPT Insight Finder. It retrieves data based on the selected views.

To configure the model 

  1. Log in to BMC Helix Innovation Studio.

  2. On the Workspace tab, from the Application launcher, select HelixGPT Agent Studio.
  3. Click Visit deployed application.
  4. Click Settings.
  5. On the Settings page, select HelixGPT > Connections > AI service providers.
    The details of the supported AI service provider models are displayed in the card format.
    Use the toggle button to switch from Card View to Table View.
  6. Click the required model to open the model configuration.
  7. In the Edit model window, specify the following details:
    FieldDescription
    API endpoint URLThe URL of the API endpoint.
    API keyThe API key required for authentication.
    Default configThe additional configurations required for a specified vendor.
  8. (For API type authentication) After specifying all the required fields, click Test Connection to test the AI service provider settings. If the connection test fails, verify that you have entered the correct details.
  9. Save the changes.

To create sub-agents 

Create sub-agents by copying the configuration details of the existing out-of-the-box sub-agents.

  1. Log in to BMC Helix Innovation Studio.

  2. On the Workspace tab, from the Application launcher, select HelixGPT Agent Studio.
  3. Click Visit deployed application.
  4. Click Agents.
  5. Select the out-of-the-box sub-agent that you want to copy:
    - For BMC Helix ITSM: Select ITSM Query Builder.
    - For BMC Helix AIOps: Select AIOps Chat Agent.
    - For BMC Helix Vulnerability Resolver: Select Vulnerability Query Builder Agent.
  6. Click Add agent.
  7. On the Add agent page, specify the following details by referring to the configurations of the out-of-the-box sub-agent that you selected in step 5:
    FieldDescription
    NameSpecify a unique name for the agent.
    TypeSelect the same type as that of the out-of-the-box sub-agent.
    ToolsSelect the same value as that of the out-of-the-box sub-agent. 
  8. Save the changes.
    The agent is created.
  9. From the Agents list, select the newly created agent.
  10. (Optional) On the Model tab, select the model and provider and click Apply.
    The changes are saved, and a confirmation message is displayed.
  11. Repeat these steps to create sub-agents for other products.

To create an Insight Finder Supervisor Agent

  1. Log in to BMC Helix Innovation Studio.

  2. On the Workspace tab, from the Application launcher, select HelixGPT Agent Studio.
  3. Click Visit deployed application.
  4. Click Agents.
  5. Select Insight Finder Supervisor Agent.
  6. On the Add agent page, specify the following details by referring to the configurations of the out-of-the-box sub-agent:
  7. FieldDescription
    NameSpecify a unique name for the agent.
    TypeSelect Insight Finder Supervisor Agent.
    Sub-agentsSelect the sub-agents that you created for the products you are using.
    Model(Optional) Select the model and provider.
  8. Retain the default configurations for the other fields.
  9. Save the changes.
    The agent is created.
  10. Copy the code block from the configuration section of the out-of-the-box agent:
    1. Log in to BMC Helix Innovation Studio.
    2. On the Workspace tab, from the Application launcher, select HelixGPT Agent Studio.

    3. Select the Agent checkbox, and click Edit data.

    4. Search for Insight Finder Supervisor Agent.

    5. Under the Display ID column, click the value corresponding to Insight Finder Supervisor Agent.
      The agent configuration details are displayed.

      View the configuration parameters of the Insight Finder Supervisor Agent

    6. Copy the code block from the Configuration section.

      Code block

      {
          "recursion_limit": 25,
          "sub_agents": {
              "itsm_query_builder": {"key": "ITSM Query Builder"},
              "aiops_query_builder": {"key": "AIOps Chat Agent"},
               "vulnops_query_builder":{"key":"Vulnerability Query Builder Agent"}
          }
      }

    7. Search for the Insight Finder Supervisor Agent that you created.

    8. Under the Display ID column, click the value corresponding to the agent.

    9. In the Configuration section, paste the code block that you copied earlier.

    10. Save the changes.

To create a skill

  1. Log in to BMC Helix Innovation Studio.

  2. On the Workspace tab, from the Application launcher, select HelixGPT Agent Studio.
  3. Click Visit deployed application.
  4. On the Skills tab, perform these steps:
    1. From the Application list, select BMC Helix Dashboards.
    2. Click Add skill, and specify these details:
      FieldDescription
      NameSpecify a unique name for the skill.
      ProviderSelect a provider.
      ModelSelect a model.
      Instruction typeSelect Agent.
    3. Click Add.
      The skill is created, and a confirmation message is displayed.
  5. Select the newly created skill.
  6. Click the Agent tab.
  7. Search for the newly created Insight Finder Supervisor Agent and select it.
  8. Click Apply.
    The agent is added to the skill, and a confirmation message is displayed.

Important
HelixGPT Insight Finder uses the model configured for the skill. To use a different model (supported by HelixGPT Insight Finder), update the skill configuration. However, if a model is configured for the sub-agent, HelixGPT Insight Finder uses this model instead of the model configured for the skill.

To configure the skill ID 

A skill indicates a capability to perform specific tasks, such as generating responses. To enable HelixGPT Insight Finder with the required skill, add the skill ID of the Insight Finder Supervisor skill to the agent.

  1. Obtain the ID of the Insight Finder Supervisor skill by performing the following steps:
    1. Log in to BMC Helix Innovation Studio.
    2. On the Workspace tab, from the Application launcher, select HelixGPT Agent Studio.
    3. Click Visit deployed application.
    4. From the Application list, select BMC Helix Dashboards.
    5. Copy the ID from the ID column corresponding to the Insight Finder Supervisor Skill that you created.

      The model ID of the Insight Finder Supervisory skill that you created.
       
  2. Log in to BMC Helix Portal and open BMC Helix Dashboards.
  3. From the navigation meu, select HelixGPT Insight Finder > Settings.
  4. In the HelixGPT Skill identifier field, enter the skill ID that you copied earlier.
  5. Save the changes.

To use specific views in HelixGPT Insight Finder 

HelixGPT Insight Finder supports a few out-of-the-box views and the custom views that you create based on supported out-of-the-box views. You can restrict the scope of HelixGPT Insight Finder by selecting specific out-of-the-box or custom views. It displays data only from the selected views.

Important
The custom views that you already created might not work. We recommend creating new custom views based on the supported out-of-the-box views.

  1. From the navigation menu, select HelixGPT Insight Finder > Configure views.
    A list of supported and unsupported out-of-the-box and custom views are displayed.

    View a list of supported and unsupported custom views.
     
  2. On the Supported views tab, review the list of supported views and create a custom view by cloning the required out-of-the-box view. Perform the following steps:
    1. Open Reporting Metadata Studio.
    2. Import the required out-of-the-box view and customize it.
    3. Publish the customized view.
      The newly created custom view is added to the Supported views tab. For more information about creating custom views, see Creating customized views.
  3. Select the checkbox corresponding to the newly added custom view.
  4. (Optional) If you want HelixGPT Insight Finder to stop using the default out-of-the-box view based on which you created the custom view, clear the checkbox corresponding to the out-of-the-box view.
  5. Click Submit.
    The changes are saved, and a confirmation message is displayed.

 

Tip: For faster searching, add an asterisk to the end of your partial query. Example: cert*