Amazon Quick's Dataset Q&A cuts query time from 90 min to under 5 min

Amazon Quick’s Dataset Q&A bridges the gap between a business question and a trustworthy answer. Instead of waiting hours for an analyst, users ask in plain English—e.g., “How is churn trending for this product?”—and the system generates SQL across millions of rows with no sampling. It applies existing row-level and column-level security policies automatically, preserving governance. The Explanations feature surfaces the full reasoning chain: tools invoked, SQL generated, filters applied, and assumptions made. For BI engineers, this turns weeks of iterative testing into focused curation sessions. In a real-world deployment with AWS’s Technical Field Communities program, Dataset Q&A improved query accuracy by over 48% and slashed resolution time from 90 minutes to under 5 minutes across 15,000 members.

Semantic Enrichment tackles the deeper challenge of business vocabulary. A column named “revenue” might mean gross, net, pre-returns, or accrual—without context, even the best model guesses. Analysts now teach AI their domain’s semantics during metadata configuration, so SQL queries reflect actual business definitions, not column-name guesswork. Together, these capabilities let data professionals move from ticket-driven analytics to instant, governed self-service. Amazon Quick positions itself as the platform that scales AI-powered decisions from small datasets to enterprise-wide data lakes with tens of millions of rows, all while maintaining reproducibility and trust.