Software

Amazon Introduces Bedrock AgentCore Browser Tool for Legacy App Automation

Amazon Bedrock AgentCore Browser Tool helps automate legacy web applications, addressing challenges faced by enterprises in healthcare, manufacturing, and finance. A large insurance company processes tens of thousands of policy changes annually using legacy systems.

Image: AWS Machine Learning

Amazon Bedrock AgentCore Browser Tool enables enterprises to automate legacy web applications that require human-like interaction. Traditional Robotic Process Automation (RPA) struggles to scale these tasks, creating significant cost centers and barriers to digital transformation. The tool uses Playwright integration through WebSocket-based Chrome DevTools Protocol (CDP) connections, allowing AI agents to interact with legacy web interfaces securely and efficiently. This solution addresses the complexities of legacy systems that lack modern REST APIs and rely on server-side middleware to generate HTML for browsers. Organizations can now implement sophisticated automation workflows that scale from single-step automations to complex multi-agent workflows, enhancing operational efficiency while maintaining compliance and security. Source: awsml

The tool runs a managed Chromium instance in the cloud, accessible over HTTP or HTTPS, regardless of the original browser design. Combined with Strands Agents for model-driven orchestration, the solution integrates with Amazon Bedrock foundation models (FMs) through Amazon Bedrock AgentCore runtime, providing session-isolated security with AWS Identity and Access Management (IAM) controls and full audit trails. This allows companies to modernize critical workflows while supporting regulatory compliance requirements and preserving human oversight. The reference implementation demonstrates an AI-powered digital worker built with Amazon Bedrock AgentCore Browser Tool and Strands Agents, covering architecture, design decisions, and an enterprise blueprint with terraform deployment. The complete source code is available on GitHub. Source: awsml

Legacy web applications often face integration complexity, compliance requirements, and scalability limitations. These systems, built decades ago on server-side middleware, generate HTML, CSS, and JavaScript for browsers without modern REST APIs. Automation solutions must interact through web interfaces designed for human users, featuring complex workflows, dynamic form validation, and session-dependent state management that traditional RPA bots cannot reliably navigate. Authentication mechanisms like multi-factor authentication (MFA) and proprietary single sign-on (SSO) further complicate automation, leading to a long tail of manual interventions. Source: awsml

Key points

Amazon Bedrock AgentCore Browser Tool automates legacy web applications that require human-like interaction.
A large insurance company processes tens of thousands of policy changes annually using legacy systems.
The tool uses Playwright integration through WebSocket-based Chrome DevTools Protocol (CDP) connections.
Amazon Bedrock AgentCore Browser Tool runs a managed Chromium instance in the cloud accessible over HTTP or HTTPS.
The solution integrates with Amazon Bedrock foundation models (FMs) through Amazon Bedrock AgentCore runtime.
Legacy web applications often face integration complexity, compliance requirements, and scalability limitations.
Authentication mechanisms like multi-factor authentication (MFA) and proprietary single sign-on (SSO) further complicate automation.

Source: AWS Machine Learning Read the original →

WRITTEN BY

Theo Almeida

AI Software & Developer Tools

Theo covers AI software, developer tools, frameworks, and the platforms builders use every day.