AI-Generated SQL Semantic Layer Practices for QuickSight Chat
AWS ML Blog · rss · 2026-07-08
This article explains how to configure multi-dataset Topics for Amazon QuickSight Chat, leveraging generative AI to dynamically generate complex SQL at query time, thereby removing the constraints of pre-defined joins.
Comparison of Query Modes:
- Defined Relationships Mode: Pre-builds a Directed Acyclic Graph (DAG) supporting only inner joins, suitable for managed reporting.
- AI-Generated SQL Mode: Based on intent parsing, supports complex logic like outer joins, unions, and subqueries, ideal for exploratory analysis.
Semantic Guidance Stack:
To reduce AI ambiguity in interpreting data, the article proposes a seven-layer metadata framework:
- Dataset-level instructions: Explains table purpose, granularity, and primary keys.
- Topic-level instructions: Defines cross-table logic and disambiguation rules.
- Synonyms: Maps business vocabulary to technical fields.
- Field descriptions: Explains column meanings and valid ranges.
- Column exclusions: Hides internal noise fields.
- Calculated fields: Pre-defines common business metrics.
By refining these metadata elements, you can effectively guide the AI to write accurate SQL queries that align with business logic.
More from Apps
- Strangeworks launches Aura to turn enterprise ops into production optimization systems — whurley · 2026-07-22
- OpenAI rolls out voice in GPT-Live, but the UI obscures search and reasoning — Graham_dePenros · 2026-07-22
- Meta AI adds interleaved image-and-text input to its text box — ezyang · 2026-07-22
- Rendergeist Pulse adds text overlays, saved presets and higher-quality exports — bennash · 2026-07-22
- EmblemVault adds limits, stops and multi-entry via DefinitiveFi Flash on Solana and EVM — adamamcbride · 2026-07-22
- A Reddit demo argues online stores should expose carts and pricing through MCP — gelembjuk · 2026-07-22