01
Discover
Describe one approved business use case. The Agent searches catalog metadata and recommends relevant tables inside the pinned source boundary.
Independent guide · Preview · Source checked August 4, 2026
Turn catalog metadata into a reviewed Quick context boundary
The Amazon Quick Agentic Catalog Experience helps curators discover assets, create DirectQuery Datasets and Topics, and inherit selected semantics from AWS Glue or Databricks. The Agent accelerates the workflow; authors still own every table, relationship, permission, and release decision.
Curator flow
Four assisted stages, with review at every boundary
02
Create
Review the recommendations, then create a focused set of Catalog-Generated Datasets as DirectQuery representations.
03
Relate
Review inherited and inferred relationships, cardinality, grain, and known totals before creating a multi-dataset Topic.
04
Inherit
Use upstream table and column descriptions as read-only semantics, then govern manual sync and any later customization.
Current boundaries
Preview automation does not collapse data governance
Supported preview catalogs
AWS Glue Data Catalog · Databricks Unity Catalog
Generated data path
DirectQuery by default; source data is queried at its origin
Semantic authority
The upstream catalog remains the source of truth
Author control
Review tables, descriptions, inherited and inferred relationships before proceeding
Identity propagation
Optional and DirectQuery-only; Glue TIP or Databricks OAuth 3LO
Customization consequence
SPICE or transformations end semantic sync and propagation
Decision guides
Follow the guide for the control you are changing
Setup guide
Connect a catalog, scope discovery, review recommended assets, create Datasets and a Topic, and validate before sharing.
Glue vs Databricks
Compare the two supported preview catalogs by authoritative source, query path, authentication, identity, and operating ownership.
Identity propagation
Separate connection auth from per-user query enforcement, DirectQuery, Glue trusted identity, Databricks 3LO, RLS, and CLS.
Datasets and Topics
Understand generated asset behavior, inherited descriptions, relationship review, focused context, Q&A, and dashboard validation.
Semantic sync
Govern inherited metadata, manual sync, source authority, Dataset forks, drift detection, and the limits of Agent recommendations.
Agentic vs manual
Choose a catalog representation, custom Dataset, or deliberate hybrid from semantics, transformations, identity, and lifecycle ownership.
Amazon Quick catalog questions, answered
It is a preview, AI-assisted curator workflow for discovering catalog assets, bulk-creating DirectQuery Datasets, assembling Topics, and inheriting selected upstream semantics from AWS Glue Data Catalog or Databricks Unity Catalog.
No. AWS describes Quick as a consumer of upstream catalog metadata. The catalog remains authoritative, and generated representations can manually sync inherited definitions.
No. AWS requires authors to review discovered tables, inherited descriptions, and inferred relationships. Reconcile joins, grains, metrics, permissions, and known results before sharing.
No. It is optional and additive. It applies only to DirectQuery Datasets; otherwise Quick-managed row-level and column-level security are documented alternatives.
No. This is an independent, source-checked guide. It does not connect a catalog, access AWS or Databricks data, create Datasets, enforce permissions, or establish a Flowith integration.
This independent guide is not an Amazon Quick console or Flowith integration. It does not connect catalogs, access data, create assets, or enforce permissions. Verify preview status, supported catalogs, authentication, DirectQuery, inherited metadata, identity, Regions, sharing, and service terms before use.