AWS shows how Amazon Quick connects tables and documents
Its walkthrough brings both data types into one conversational assistant through a shared workspace.
AWS’s walkthrough combines documents indexed from S3 storage with a tabular dataset enriched with descriptions. A Space connects these sources to a custom chat agent. Questions about structured data become SQL database queries; the assistant can combine results with document findings and citations. Following the example requires an Enterprise subscription, administrator-granted bucket access and permissions to create the components.
The workspace documentation distinguishes sharing from data access. Linking an existing resource preserves its permissions: sharing the Space does not automatically grant access to that resource. Files uploaded directly into the Space, however, are available to everyone with access to it. Upload status matters too: Text ready permits queries about text while images and tables still process; Ready means text and media have been processed.
The agent guide separates previewing, saving and sharing. After changing settings, Update preview saves them in the preview before testing. Leaving creation without Launch deletes the preview; launching saves the agent in the library but keeps it private until shared. For a linked agent, answers draw on model knowledge and linked resources, with one additional temporary knowledge source allowed during chat. Later launches of edits to a shared agent update that same agent for its existing users.
