March 11th, 2025
We are excited to announce that knowledge bases deployed by BlueXP workload factory now support hybrid search and re-ranking, significantly improving the relevance and accuracy of search results from your organization's data.
While semantic search is helpful to understand the intent behind user queries and provides relevant results even if keywords in user queries don’t match, it may fall short for scenarios where data sources include heavy use of specific terms and jargon such as product codes, names, abbreviations, etc. Keyword based search excels at exact matches, however, does not understand the context or meaning of the query. Hybrid search combines the strengths of traditional keyword-based search with advanced dense vector based semantic search techniques to provide users with more relevant and accurate information. The results from both hybrid and semantic search are then combined and refined further by using advanced re-ranking models such as Cohere Rerank 3.5 and Amazon Rerank 1.0 to return the most relevant results to the users.
This enhanced capability is now available in BlueXP workload factory for GenAI by default for newly created knowledge bases. We encourage you to explore how hybrid search and re-ranking can improve relevance of the results for your generative AI applications.
To learn more, visit the Workload Factory product page, documentation. To get started, sign up for Workload Factory.