Prerequisites to Using Augmented AI - Amazon SageMaker

Prerequisites to Using Augmented AI

Amazon A2I uses resources in IAM, SageMaker, and Amazon S3 to create and run your human review workflows. You can create some of these resources in the Amazon A2I console when you create a human review workflow. To learn how, see Tutorial: Get Started in the Amazon A2I Console.

To use Amazon A2I, you need the following resources:

  • One or more Amazon S3 buckets in the same AWS Region as the workflow for your input and output data. To create a bucket, follow the instructions in Create a Bucket in the Amazon Simple Storage Service Console User Guide.

  • An IAM role with required permissions to create a human review workflow and an IAM user or role with permission to access Augmented AI. For more information, see Permissions and Security in Amazon Augmented AI.

  • A public, private, or vendor workforce for your human review workflows. If you plan to use a private workforce, you need to set one up ahead of time in the same AWS Region as your Amazon A2I workflow. To learn more about these workforce types, see Create and Manage Workforces.


    To learn about the compliance programs that cover Amazon Augmented AI at this time, see AWS Services in Scope by Compliance Program. If you use Amazon Augmented AI in conjunction with other AWS services (such as Amazon Rekognition and Amazon Textract), note that Amazon Augmented AI may not be in scope for the same compliance programs as those other services. You are responsible for how you use Amazon Augmented AI, including understanding how the service processes or stores customer data and any impact on the compliance of your data environment. You should discuss your workload objectives and goals with your AWS account team; they can help you evaluate whether the service is a good fit for your proposed use case and architecture.