Revolutionizing Business Operations with FAST: Deploying Generative AI Agents Using Amazon Bedrock AgentCore

Introduction to Agentic Applications and Amazon Bedrock

The FAST Solution: Overview and Architecture

Deploying FAST: A Step-by-Step Walkthrough

Customizing Your Agentic Application

Leveraging AI-Assisted Development with FAST

Exploring Use Cases Beyond Chat: From Document Analysis to Workflow Automation

Conclusion: Transforming Ideas into Reality with FAST

About the Authors: Expertise Behind the Innovation

Harnessing Generative AI: Transforming Business Operations with Amazon Bedrock AgentCore

The landscape of business operations is evolving, thanks in large part to generative AI and its agentic applications. From sophisticated customer support bots to intelligent research assistants, these technologies are helping teams move swiftly from prototypes to production-ready solutions. One significant advancement in this field is Amazon Bedrock AgentCore, introduced by AWS last year. This development platform streamlines the process of building, deploying, and scaling AI agents in a production environment, giving organizations the tools they need to innovate and excel.

Introducing Amazon Bedrock AgentCore

Amazon Bedrock AgentCore offers essential building blocks such as runtime hosting, memory, tool integration, and observability, all backed by enterprise-grade security and dynamic scaling capabilities. This robust framework empowers developers to focus on creating value rather than getting bogged down by infrastructure concerns.

A standout feature is the Fullstack AgentCore Solution Template (FAST), a ready-to-deploy starter project designed to illustrate how the various components of AgentCore interact seamlessly. FAST combines AgentCore Runtime, Gateway, Memory, and Code Interpreter with a React frontend and Amazon Cognito for authentication, all defined using the AWS Cloud Development Kit (AWS CDK).

Why FAST?

FAST serves as a complete reference architecture, showing how different components work together to establish a working chat application. By utilizing this template, organizations can expedite their development processes, minimizing the time from ideation to deployment.

In this blog post, we’ll walk you through how to deploy FAST to your Amazon Web Services (AWS) account, understand its structure, and extend it to fit your specific needs.

Solution Overview

FAST delivers a full-stack architecture specifically tailored for deploying agents on Amazon Bedrock AgentCore. The template streamlines various tasks:

• Authentication: Managed securely through Amazon Cognito.
• Frontend Hosting: A React application utilizing Tailwind CSS for a responsive UI, hosted on AWS Amplify Hosting.
• Agent Runtime: Hosted on Amazon Bedrock AgentCore, ensuring performance and scalability.
• Observability: Offers monitoring features to inspect agent performance and behavior.
• Tool Integration: Built-in support for Model Context Protocol (MCP) tool integration.

Architecture Details

At the heart of FAST is the Amazon Bedrock AgentCore Runtime, which hosts your agents. User authentication is facilitated via Amazon Cognito, securing four key integration points:

1. User sign-in to the frontend application hosted on Amazon CloudFront.
2. Token-based authentication allowing the frontend to access the AgentCore Runtime.
3. Token-based access for agents communicating with the AgentCore Gateway.
4. Authentication for API requests directed at the Amazon API Gateway.

The design is modular, enabling developers to swap components as needed, including the frontend framework or the identity provider, without affecting the core functionalities.

Documentation and AI-Assisted Development

FAST comes with comprehensive documentation, making it easier for developers and AI coding assistants alike. The repository includes:

• Steering Documents: Guidelines that AI assistants can follow.
• Feature Guides: In-depth documentation on integration patterns.
• Component Context: READMEs distributed throughout the codebase for easy navigation.

This dual approach ensures that both AI-assisted development and traditional coding are equally supported.

Deployment Walkthrough

Ready to try it out? Here’s a step-by-step guide to deploying FAST:

Step 1: Clone the Repository

git clone https://github.com/awslabs/fullstack-solution-template-for-agentcore.git
cd fullstack-solution-template-for-agentcore

Step 2: Configure Your Deployment
Edit infra-cdk/config.yaml to customize your deployment settings.

Step 3: Deploy the Backend with CDK

cd infra-cdk 
npm install 
cdk bootstrap 
cdk deploy

Step 4: Deploy the Frontend

cd .. 
python scripts/deploy-frontend.py

Step 5: Create an Amazon Cognito User
Follow the instructions to create users either manually or automatically.

Step 6: Test the Application
After deployment, access the app using the URL provided and log in with your credentials.

Customize Your Application

The baseline chat application can be modified to create an array of agentic solutions—whether it’s a document analysis tool or an integrated company workflow automation agent. The flexibility of FAST allows for easy adjustments and components swapping as requirements evolve.

Beyond Chat: Exploring More Use Cases

FAST’s architecture is not limited to chat interfaces. Developers have the freedom to adapt components for diverse applications, whether it’s document analysis, workflow automation, or monitoring services like Slack channels.

Clean Up

When you’re finished testing, you can remove all the resources created by FAST using:

cd infra-cdk 
cdk destroy --force

Conclusion

Using FAST can dramatically reduce your time to market, enabling you to build and deploy agent applications in under 30 minutes. With built-in capabilities for authentication, secure hosting, and tool integration, developers can focus on innovation rather than infrastructure.

To explore FAST further, clone the repository, deploy to your AWS account, and start customizing your agent-driven applications today. The future of business operations is here, and with tools like Amazon Bedrock AgentCore, it’s more accessible than ever.

For expert guidance, don’t hesitate to connect with the AWS Generative AI Innovation Center or leverage AWS Professional Services.

About the Authors

This post is authored by a team of experts from the AWS Generative AI Innovation Center, dedicated to pushing the boundaries of what is possible with AI in operational contexts.

By leveraging the capabilities of generative AI and frameworks such as Amazon Bedrock AgentCore, businesses can not only increase efficiency but also unlock new opportunities for innovation and growth.