Configuring clarification to improve response accuracy

Use the Clarification option to improve the accuracy of responses. You can use the following two methods for presenting the clarification option to users:

• Static clarification: Administrators can configure record definitions as predefined questions and options. These options are presented to users when additional context is needed to refine the query. This is useful for common or expected ambiguities. For more information about the configuration, see Configuring QueryClarification record.
• Dynamic clarification: Administrators can use the Knowledge Clarification prompt to dynamically generate clarification options based on the referenced documents from connected data sources. The prompt helps in scenarios where predefined questions are ambiguous.

The knowledge clarification prompt generates options at runtime based on knowledge context, whereas the QueryClarification record uses predefined entries stored in a custom record definition.

The clarification is triggered when a user's query is ambiguous and could match multiple documents from the configured data sources. Static and dynamic clarification methods are independent. You can use either or both based on your use case. If both static and dynamic clarifications are set up:

• BMC HelixGPT first checks for predefined questions in the QueryClarification records. If a match is found, it shows these questions as options to help the user clarify their intent.
• If no predefined questions are available, BMC HelixGPT uses the Knowledge Clarification prompt. It retrieves relevant documents, extracts key information, and generates clarification options. These options are presented as clickable buttons to guide the users to refine their input and get more accurate responses.
  

To configure the Knowledge Clarification prompt

Use the knowledge Clarification prompt for dynamic clarification. For more information about the prompt, see the Knowledge Clarification prompt. And for more information about configuring the prompt, see Creating and managing prompts.

To configure the predefined clarification options

1. Log in to BMC Helix Innovation Studio.
2. On the Workspace tab, click HelixGPT Agent Studio.
3. From the Records tab, select QuerClarification and click Edit data.
4. In the Data editor (QueryClarification) window, click +New.
5. In the New record window, specify the following fields:

   Field name	Description
   Clarification Question	The clarification response presented to the user.
   Facet	A category or tag used to group or filter the clarification context.
   Question	The trigger phrase or keyword entered by the user that activates the clarification record.
   Options	A JSON list of selectable clarification options presented to the user. For example:  ["option1", "option2"]
   Status	Status of the record. Indicates whether the record is active or inactive state.

6. Click Save.
7. (Optional)The threshold value for query similarity can be configured from the Configure tab of the Skill.
   • Add the threshold value as a key-value pair in the configuration. 
     For example:  "clarificationQuestionThreshold": <threshold_value>
   • The default and the recommended threshold value is 0.9; you can modify it as per your requirements.

Related topic

Creating and managing skills