Azure AISearch supports the following extended search options:
- similarity: Standard vector similarity search (default option).
- similarity_hybrid: Combines the strengths of full-text search and vector search to provide the best ranking performance.
- semantic_hybrid: Similar to the previous option plus vector search with semantic ranking. Note that semantic ranking capabilities are only available in the Basic and higher pricing tiers.
To use this advanced configuration, you must add the following parameters to the Profile Metadata.
{
"chat": {
"retriever": {
"searchType": "similarity | similarity_hybrid | semantic_hybrid"
}
}
}
Note: Assigning the semantic_hybrid value to the searchType parameter requires recreating the Index. It's recommended to create a new RAG assistant for this purpose.
When using metadata filters, ensure the metadata complies with Azure AI Search restrictions described here, especially for document key fields, which cannot contain spaces.
Hybrid search
Outperforming vector search with hybrid retrieval and ranking capabilities