What's New

Amazon Quick has launched Dataset Q&A, a conversational analytics capability that allows users to ask natural language questions directly against enterprise datasets. This feature complements the existing Dashboard Q&A functionality by enabling ad-hoc, open-ended data exploration beyond pre-built dashboards. Dataset Q&A is powered by a text-to-SQL agent that interprets questions, identifies relevant data, and generates optimized SQL in a single conversational step, all while enforcing existing governance policies such as Row Level Security (RLS) and Column Level Security (CLS).

How It Works

• At its core, Dataset Q&A uses a text-to-SQL agent that translates natural language queries into engine- and dialect-aware SQL optimized for the target data source — including Amazon Redshift, Amazon Athena, Aurora PostgreSQL, Apache Iceberg tables in S3 table buckets, and SPICE.
• Data owners can enrich datasets with custom instructions, business definitions, and field descriptions either inline within Amazon Quick or via file uploads.
• These semantic annotations, combined with dataset metadata, are ingested into a knowledge graph that models the meaning and relationships across data assets.
• An orchestrator layer uses this knowledge graph to identify the most relevant datasets and generate accurate SQL at query time.
• The system supports a wide range of analytical question types — trend analysis, time-series comparisons, ranking, multi-condition queries, and exploratory questions.
• An "Explain" capability surfaces the reasoning chain, the generated SQL, and the underlying logic so users can validate correctness before acting on results.

Why It's Important

• Dataset Q&A significantly lowers the barrier to data access for non-technical business users by eliminating the need to write SQL or rely on pre-built dashboards for every analytical question.
• Because it inherits and enforces existing RLS and CLS governance policies, organizations can safely expose this capability to a broad user base without compromising data security or compliance.
• The knowledge graph approach means that semantic context — business terminology, field relationships, and domain-specific definitions — is preserved and reused across queries, improving accuracy and reducing the need for repeated prompt engineering.
• The Explain feature also builds user trust by making the AI's reasoning transparent and auditable, which is critical for enterprise adoption of AI-driven analytics.

How It's Different

• Prior to Dataset Q&A, Amazon Quick's natural language capability was primarily scoped to Dashboard Q&A, which allowed users to ask questions within the context of an already-built dashboard and its pre-defined visuals.
• Dataset Q&A removes that constraint, enabling direct, schema-level interrogation of raw datasets without requiring a dashboard to exist first.
• Compared to generic text-to-SQL tools or standalone LLM integrations, Dataset Q&A is tightly integrated with Quick's governance layer, ensuring security policies are automatically applied rather than bolted on.
• The knowledge graph enrichment model also differentiates it from simpler prompt-based SQL generators, as it captures persistent semantic context that improves query accuracy over time and across users — rather than relying solely on the LLM's in-context understanding of schema metadata.

When to Prefer It

• Dataset Q&A is the right choice when business users need to perform ad-hoc, exploratory analysis against enterprise datasets without waiting for a BI developer to build or modify a dashboard.
• It is particularly well-suited for organizations with large, complex data estates where users may not know which dataset to query or how fields relate to one another — the knowledge graph handles dataset discovery and disambiguation automatically.
• Teams that have already invested in defining RLS/CLS policies in Amazon Quick will benefit immediately, as those controls are enforced without additional configuration.
• For use cases requiring auditability of AI-generated insights — such as financial reporting, compliance analytics, or executive decision-making — the Explain capability makes Dataset Q&A preferable to black-box alternatives.
• It is less appropriate when highly complex, multi-step analytical workflows require custom SQL logic that exceeds what a text-to-SQL agent can reliably generate, or when the underlying datasets lack sufficient semantic enrichment to guide accurate query generation.

Availability

• Dataset Q&A is generally available (GA) as of May 4, 2026, and is accessible in all AWS Regions where Amazon Quick is currently supported.
• No preview or limited-access period is indicated — the feature is fully released for production use.
• Supported data sources include Amazon Redshift, Amazon Athena, Aurora PostgreSQL, Apache Iceberg tables stored in Amazon S3 table buckets, and SPICE.
• There are no explicitly stated regional exclusions, though availability is contingent on Amazon Quick itself being available in a given region.
• Organizations should ensure their datasets are enriched with business definitions and field descriptions to maximize query accuracy, as the quality of the knowledge graph directly influences the reliability of generated SQL.