#
AI Search (Semantic Search)
AI search uses vector embeddings to retrieve content based on meaning rather than just keywords.
#
Configuring Embeddings
- Navigate to Management > Search > AI Search.
- Add fields to the AI index (e.g.,
SupportCase.Summary). - Chunking: For long fields like
Content, enable Chunk Text to ensure the entire text is indexed within the model's context window.
#
Semantic Re-ranking
AI search can be used to re-rank the results of a keyword search, bringing the most conceptually relevant items to the top.
#
Similarity Cutoff
Adjust the added and rerank cutoff values to control how strictly the AI should match similar results.