Managing data and making it accessible for insights can be challenging, especially when dealing with large, raw datasets. Teams often need to prepare data for querying, build the right schemas, and design workflows that allow technical and non-technical users to extract information efficiently. Manually writing SQL queries adds another layer of complexity, limiting accessibility for users who may not be fluent in SQL. This is where services like Amazon Athena and Amazon Bedrock can work together to simplify the process, enabling seamless querying of structured data using natural language.
In this Cloud Lab, you’ll learn how to integrate an Amazon Bedrock agent with Amazon Athena to query datasets using plain English instructions instead of SQL. You’ll start by preparing your dataset—creating an Amazon S3 bucket, uploading raw data, and configuring an AWS Glue crawler to automatically detect the schema and make the data queryable in Athena. Once the data is ready, you’ll run your first SQL query in Athena to confirm everything is set up correctly.
Next, you’ll focus on enabling natural language querying. You’ll create resources for an action group and build a Bedrock agent capable of interpreting plain English questions and converting them into SQL queries. The agent will also be able to return query results in plain English. Finally, you’ll use the agent to query the dataset directly with natural language, bypassing the need to manually craft SQL statements.
After completing this Cloud Lab, you’ll understand how to prepare and query data in Amazon Athena and how to integrate Amazon Bedrock for natural language querying. You’ll gain hands-on experience setting up the full workflow—from preparing the dataset and schema to enabling a Bedrock-powered agent that transforms natural language requests into actionable insights. This will give you a practical foundation for building intelligent, user-friendly data querying solutions in the cloud.
Below is the high-level architecture diagram of the infrastructure you’ll create in this Cloud Lab:
