# Amazon Quick brings structured and unstructured data together in a single conversational AI agent > Amazon Quick combines structured data queries with document search in a single chat agent, connecting S3-backed knowledge bases and enriched datasets under one Enterprise subscription. - Type: AI news - Page: https://completeaitraining.com/news/amazon-quick-brings-structured-and-unstructured-data/ - Published: 2026-11-04 - Updated: 2026-11-04 - Author: Jeroen Erne - Jobs: Customer support, Education, Government, Hospitality and events, IT and development, Legal, Operations - Reading time: 3 min Amazon Quick now lets organizations query structured data from datasets and unstructured data from documents in a single conversation, without switching tools or writing separate queries. The capability connects S3-backed knowledge bases with enriched datasets through a unified AI agent, available to users with an Enterprise subscription. The architecture works in three layers. The data layer holds unstructured content like PDFs and reports in Amazon S3, indexed through a knowledge base, while structured metrics and KPIs sit in enriched datasets with semantic descriptions the agent uses to translate questions into SQL. A workspace layer called a Space brings both data types together. The agent layer then uses a custom chat agent linked to that Space to determine whether each question needs the knowledge base, the dataset, or both, returning a combined response with source citations. #### How to set up the unified experience Start by creating an S3 integration from the Knowledge page in the Amazon Quick console. Provide a name and the S3 URI of the bucket holding your documents. After the integration is created, build a knowledge base that indexes the S3 content. You can index everything in the bucket or specify S3 prefixes for specific folders and files. Next, upload a structured data file such as a CSV to create a dataset. After ingestion into SPICE, Amazon Quick's in-memory engine, enrich the dataset with semantic context. Fix any auto-detected data types that are incorrect, then add a dataset-level description, custom instructions for how the AI should query the data, and column-by-column descriptions. This context helps the agent generate accurate SQL from natural language questions. Create a Space and add both the knowledge base and the enriched dataset as data sources. Then build a custom chat agent, configure its persona and communication style, and link it to the Space. The agent now has a single location to query across structured and unstructured data. #### What the agent can do Questions like "How did sales change by region in 2025?" prompt the agent to combine results from the sales dataset with documents in the knowledge base, returning regional breakdowns with citations. Asking about a specific document, such as "What are the key strategic priorities outlined in our product strategy memo?", surfaces the relevant sections from that file. A question that spans both data types - for example, "How do our 2025 sales results align with the market trends identified in our analysis report?" - pulls from CSV data and the market analysis document in one response. The agent can also generate formatted artifacts on command. Users can request PowerPoint presentations, Excel spreadsheets, or Word documents through natural language prompts, and the agent produces downloadable files. #### Security and access controls Amazon Quick operates under the AWS shared responsibility model. Administrators grant S3 bucket access with read-only permissions by default. If documents need per-user or per-group access, document-level ACLs must be enabled when the knowledge base is created - the choice cannot be changed later. Chat agents and knowledge bases remain private until explicitly shared, and responses respect the permissions of the person asking the question. Built-in content safety screening applies to chat interactions, and administrators can add up to 50 blocked words or phrases. For teams working with sensitive business documents across IT, legal, or operations functions, [AI Data Analysis Courses](https://completeaitraining.com/tag/data-analysis/) cover techniques for structuring and querying data through conversational interfaces. #### Why this matters for operations and IT teams Operations professionals, IT staff, and legal teams often spend hours pulling reports from databases, then cross-referencing them against documents stored in separate systems. Amazon Quick eliminates that manual correlation step. A single question in the chat agent can return figures from a sales dataset alongside relevant passages from a market analysis PDF, with citations for both. The agent handles the SQL generation, so business users don't need to know query languages or where each piece of data lives. For teams managing compliance documentation, policy files, and operational metrics, this means faster answers during audits, quarterly reviews, and ad-hoc analysis without involving a data analyst for every request. --- This is the Markdown form of https://completeaitraining.com/news/amazon-quick-brings-structured-and-unstructured-data/. The whole site is mapped at https://completeaitraining.com/llms.txt.