What changed
Amazon Bedrock Managed Knowledge Base now supports tabular search, allowing users to get answers from tables through agentic retrieval. This feature enables the system to look up values and run calculations across every relevant row in one or more tables or sheets when a question requires data from CSV, XLSX, or XLS files.
What this means in practice
When a question needs data from tables, tabular search combines results from the tables with facts from related documents to answer analytical questions that require multiple steps. For example, a finance team closing the quarter can determine which departments must write variance commentary by analyzing the variance policy’s threshold, the quarter-close memo’s budget revisions, and math over every department in the profit and loss (P&L) workbook. Tabular search runs code over the original file in an isolated sandbox, ensuring the answer uses the whole table instead of only the rows that retrieval returns.
Limitations and availability
Tabular search is built for questions that need calculations or lookups over table files it retrieves. Running code adds at least one more agent turn, so answers that use table data take longer than text-only answers. Governed metrics, recurring dashboards, warehouse-scale joins, and reporting controls still belong in your business intelligence (BI) tools. Tabular search works on table files stored next to your documents, and you don’t define a schema. The feature is available through the AgenticRetrieveStream request, which can search one knowledge base or several, and works with connectors like Amazon S3, Microsoft SharePoint, and Atlassian Confluence.
