Modifying the default Static Query Clarification tool settings


As an administrator, configure the Static Query Clarification tool to specify the similarity threshold and the value for the K parameter at the tool level.

Configure the K parameter to determine the number of questions that are most relevant or match the intent of the user query . 

You can configure the similarity threshold in the tool to identify the most relevant record definition that matches the user query. The record definition is then displayed to users as a clarifying question, with the available options presented as clickable buttons. 

The Static Query Clarification tool is available in HelixGPT Agent Studio, but you cannot modify it directly. However, you can add a custom tool to modify the default tool configuration.

Before you begin

Make sure you have cloned a prebuilt agent, because you can only add a custom tool to a cloned prebuilt agent.

For more information, see Cloning an agent.

To modify the default Static Query Clarification tool settings

  1. Add a custom tool in your environment. For more information, see Creating and editing tools and toolkits.
    The new tool must replicate the field value configuration of the existing prebuilt Static Query Clarification tool.
  2. In the Configuration field, modify the following settings:
    SettingDescription
    "K"

    Number of most relevant clarification records to return.

    The clarification question with options is displayed only if at least one relevant record is found. 

    clarification_question_threshold

    A key-value pair that specifies the similarity threshold.

    The similarity threshold defines how closely a user query must match an intent or knowledge article. 

    Default configuration:

    {
      "k": 1
    }
  3. Click Save.

Where to go from here

Creating and editing tools and toolkits

Configuring clarification in BMC HelixGPT to improve response accuracy

 

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

BMC HelixGPT 26.1