The Knowledge Content Search tool enables AI Agents to retrieve relevant knowledge content during task execution. Unlike the FAQ tool, which performs both content retrieval and response generation, the Knowledge Content Search tool is responsible only for retrieving relevant knowledge results.

When invoked, the tool searches the ingested knowledge base, identifies the most relevant content, and returns the top-ranked knowledge chunks to the AI Agent. The AI Agent then uses this content as context to generate a response based on the task instructions.

This retrieval-first approach helps improve response times while providing greater control over how responses are generated.

How It Works

When a task uses the Knowledge Content Search tool:

1. The AI Agent can optionally rewrite the user's query using the conversation context.
2. The query is sent to the Knowledge Content Search tool.
3. The tool performs semantic search and ranking across the configured knowledge sources.
4. Relevant knowledge chunks are retrieved and returned to the AI Agent.
5. The AI Agent uses the retrieved content to generate a contextual response.

Benefits

Using the Knowledge Content Search tool provides the following advantages:

• Improves response performance by separating retrieval from response generation.
• Retrieves relevant knowledge content for use as conversational context.
• Allows responses to be guided by task-specific instructions and prompts.
• Supports conversational context through optional query rewriting.
• Enables more flexible and customizable retrieval workflows.

Create a Knowledge Content Search Tool

1. On the Tools window, click + Add Tool.
2. On the Select Tool screen, select Knowledge Content Search, and then click Next.
3. Configure the tool settings and click Save.

Basic Information

Configure the following details:

• Name – Enter a unique name for the tool.
• Description – Provide a brief description of the tool's purpose.

​

Setup Function

The Setup Function section contains the configuration required for knowledge retrieval.

Re-word User Query for Search

Enable Re-word User Query for Search to allow the AI Agent to rewrite the user's query before performing the search.

When enabled, the AI Agent uses conversation context to generate a more detailed search query, which can improve retrieval accuracy.

When disabled, the search is performed using only the current user query.

​

Query Parameter Name

Specify the parameter name used when generating the rewritten search query.

The default value is recommended unless custom implementation requirements exist.

Query Parameter Description

Provide guidance for how the rewritten search query should be generated.

The default configuration is recommended for optimal search performance.

Add Custom Code

Optionally add custom code to enhance retrieval behavior or provide additional context to the search process.

Common use cases include:

• Restricting searches to specific topics.
• Passing additional context to the retrieval workflow.
• Customizing retrieval logic for business-specific requirements.

Query Rewriting Behavior

Query Rewriting Enabled

When query rewriting is enabled, the AI Agent uses conversation history to generate an optimized search query before retrieving knowledge content. This typically produces more relevant search results.