Working with Amazon Quick Sight Topics
| Applies to: Enterprise Edition |
| Intended audience: Amazon Quick administrators and authors |
A Topic in Quick Sight is the multi-dataset semantic layer that brings multiple enriched datasets together into a unified data model. Topics let you define relationships between datasets so that Quick Sight can perform runtime joins across them — whether you're building analysis visuals or asking natural language questions through Amazon Quick chat.
With Topics, you can:
-
Add up to 12 datasets to a single Topic and define relationships (join keys) between them.
-
Add custom instructions that guide the AI engine on cross-dataset logic, disambiguation rules, and business definitions.
-
Use the Topic as a data model in Quick Sight analysis sheets to build visuals with fields from multiple datasets — Quick Sight performs runtime inner joins automatically.
-
Use the Topic in Amazon Quick chat where the LLM-powered chat agent traverses relationships across datasets, generates cross-dataset SQL with appropriate joins, and returns unified answers.
-
Preserve each dataset at its native granularity — no pre-joining or denormalization required.
Each dataset in a Topic should be independently enriched with semantic metadata (column descriptions, synonyms, semantic types, and custom instructions) before being added to the Topic. For more information about enriching datasets, see Data Preparation Experience (New).
Note
If you have existing Topics that were created before this release, they are now classified as legacy Topics. Legacy Topics continue to work as before. For more information, see Working with legacy Topics.
To learn more about working with Topics, continue to the following sections.
