Modifying the default Knowledge tool settings

As an administrator, configure the Knowledge tool to search, retrieve, and get suggestions for relevant knowledge articles. This tool enhances issue resolution by providing accurate information in a timely manner. The Knowledge tool is a specialized tool designed for performing semantic searches by using the OpenSearch retriever.

The tool searches for documents in OpenSearch by understanding the user's intent and analyzing the context of their query, rather than just matching keywords. It helps agents respond to user queries in BMC HelixGPT based on the documents present in OpenSearch.

The tool assists the following AI agents in searching the relevant documents in OpenSearch:

• Employee Navigator
• Employee Navigator – Cards
• Employee Navigator - MS Teams
• Employee Navigator Supervisor
• Employee Navigator Supervisor - MS Teams
• Employee Navigator ServiceNow
• BWF Agent
• ITSM Agent
• Voice Agent

The Knowledge tool is available out-of-the-box, but you cannot modify it directly. However, you can add a new tool to modify the default tool configuration.

Before you begin

Make sure that you have cloned an out-of-the-box agent, because you can only add a custom tool to a custom agent—this includes both cloned out-of-the-box agents and newly created agents.

For more information, see Cloning an agent.

To modify the default Knowledge tool settings

1. Add a new tool to your environment. For more information, see Adding tool and toolkit for an AI Agent.
   The new tool must replicate the same field values configuration as the existing out-of-the-box Knowledge tool. However, the tool can set only one setting or none at all, as there is no strict requirement for all configurations to be present.
2. In the Configuration field, modify the following settings:

   Setting	Description
   k	numberOfDocumentsToReturn: Specifies the maximum number of documents to be retrieved in response to a query. This parameter helps control the result set size, ensuring that only the most relevant documents are returned based on the query's ranking criteria.
   fetch_k	numberOfDocumentsToFetchForMMR: Specifies the maximum number of similar and diverse documents to be retrieved based on the MMR algorithm.
   knn_k	numberOfNearestNeighborsToSearch: Defines the number of closely relevant documents to be evaluated during a vector search operation. This parameter determines how many closely related documents must be considered for similarity ranking, influencing the accuracy and speed of the vector search result.
   lambda_mult	degreeOfDiversity: Defines the level of variation considered while generating or selecting results. A higher value increases the diversity of responses, producing a wider range of outputs, while a lower value generates more focused and consistent results. This parameter helps balance creativity and precision in data processing or response generation.
   low_watermark_threshold	Specifies the minimum relevance threshold required for similarity_score_threshold. The value must be between 0 and 1
   search_type	Defines the type of search the Retriever should perform. Valid values: similarity—returns similar documents similarity_score_threshold—returns similar documents only if they meet the minimum threshold specified by lowWaterMarkThreshold mmr—returns similar and diverse documents based on maximum marginal relevance
   grade_documents	Indicates whether to grade documents based on a threshold.
   grade_threshold	Minimum score required for a document to be graded. Valid values range from 0 to 5.

   Default configuration:
   {
     "grade_documents": true,
     "grade_threshold": 3
   }

3. Click Save.

Where to go from here

Adding tool and toolkit for an AI Agent