Announcing the Agentic Catalog Experience in Amazon Quick

Amazon Quick has launched the Agentic Catalog Experience, an AI-powered workflow aimed at data curators who need to find and promote assets from existing enterprise catalogs. Curators can search upstream catalog holdings in natural language and have the system auto-create Datasets and Topics that carry inherited semantics from the source. The feature is in preview against two major catalog backends: AWS Glue Data Catalog and Databricks Unity Catalog.

On the product mechanics side, the experience sits between the upstream catalog and Quick’s own Dataset and Topic objects. Natural-language discovery replaces hand-browsing of tables, databases, and related metadata. When a curator selects or accepts an asset, Quick generates Dataset and Topic definitions and pulls semantics forward from the catalog rather than requiring a blank-slate remodel. Supporting both Glue and Unity Catalog means the same curator workflow can span AWS-native and Databricks-managed estates without a separate tool path for each.

For engineers and builders, the value is less about chat novelty and more about cutting the manual gap between catalog registration and analytics-ready objects. Teams that already invest in Glue or Unity Catalog can treat that investment as the semantic source of truth and let Quick materialize consumer-facing Datasets and Topics with less copy-paste modeling. Data platform owners who standardize naming, types, and business terms in the catalog get a clearer path for those terms to show up in Quick without a second curation pass.

In market terms, the announcement ties Quick more tightly to the two catalog systems many enterprises already run as system of record for tables and governance metadata. Dual preview coverage of Glue and Unity Catalog positions the Agentic Catalog Experience for hybrid stacks rather than a single-vendor catalog lock-in. That matters for organizations that keep warehouse and lake workloads split across AWS and Databricks but want one curator-facing path into Quick.

Practical takeaway: if you curate for Quick and already publish into Glue or Unity Catalog, evaluate the preview on a small set of well-governed upstream assets and check whether auto-created Datasets and Topics preserve the semantics you care about—descriptions, classifications, and related business context—without manual repair. Watch how far natural-language discovery works on messy or sparsely documented catalogs, and whether inherited semantics stay accurate enough for production Topics before you expand beyond pilot domains.