​
What it does

The Search Knowledge Base tool allows agents to search through your organization’s knowledge bases to find relevant information. It uses AI-powered semantic search combined with traditional text search for comprehensive results.
For complete information about Knowledge Bases, including setup, file upload, and management, see Knowledge Bases.

​
Key features

  • Semantic Search: AI understands context and meaning, not just keywords
  • Hybrid Results: Combines vector similarity with traditional text search
  • File Filtering: Include or exclude specific documents from search
  • Ranked Results: Results ordered by relevance with similarity scores
  • Multi-Format Support: Search across PDFs, Word docs, spreadsheets, and text files

​
Parameters

ParameterTypeRequiredDescription
querystringYesThe search query to find relevant information in knowledge bases
knowledge_base_idintegerYesThe ID of the specific knowledge base to search in
top_kintegerNoNumber of results to return (default: 10)
include_file_idsarrayNoOptional list of file IDs to restrict search to only these documents
exclude_file_idsarrayNoOptional list of file IDs to exclude from search results

​
Common use cases

​
Find policy information

query: "employee vacation policy"
knowledge_base_id: 123
top_k: 5
Search company policies and procedures for specific information.

​
Research financial data

query: "quarterly revenue trends"
knowledge_base_id: 456
top_k: 10
Find financial insights from uploaded reports and documents.

​
Search specific documents

query: "budget analysis"
knowledge_base_id: 123
include_file_ids: [45, 67, 89]
top_k: 8
Focus search on recently uploaded or relevant documents only.

​
Exclude outdated content

query: "product specifications"
knowledge_base_id: 123
exclude_file_ids: [12, 34]
top_k: 10
Search while excluding old or deprecated documents.

​
Technical documentation lookup

query: "API authentication methods"
knowledge_base_id: 789
top_k: 5
Find specific technical information from documentation collections.

​
Best practices

  • Use specific, descriptive queries for better results
  • Start with broader searches, then use file filtering to narrow down
  • Monitor similarity scores to gauge result quality
  • Organize knowledge bases by topic for more focused searches
  • Regularly update knowledge bases by removing outdated documents

​
Troubleshooting

“No results found”
  • Try broader or alternative search terms
  • Check that the knowledge base contains relevant documents
  • Verify the knowledge base ID is correct
“Knowledge base not found”
  • Confirm the knowledge base ID exists in your organization
  • Check that you have access to the specified knowledge base
  • Ensure the knowledge base has been successfully created
“Low similarity scores”
  • Try rephrasing your query with different keywords
  • Use more specific terms related to your content
  • Check that your documents contain the information you’re searching for
“File filtering not working”
  • Verify that the file IDs exist in the knowledge base
  • Check that include/exclude file ID arrays are properly formatted
  • Ensure files haven’t been deleted from the knowledge base
  • Call API - Fetch external data to supplement knowledge base searches
  • Execute Python - Process and analyze search results
  • Send Email - Share search results with team members