AWS AgentCore: Automate Legacy Apps with AI

AWS Lets AI Agents Click Through Old Web Apps That Have No API

Short answer

AWS released the Bedrock AgentCore Browser Tool, a managed service that lets AI agents navigate and operate legacy web applications through the browser itself, clicking, filling forms, reading screens, without an API. For companies running on old internal tools or vendor portals with no integration path, this opens a route to automate workflows that were previously stuck requiring manual data entry.

What this means for operators

Most 10-200 person B2B companies carry at least one legacy system with no API: an old order-management tool, a supplier portal, an internal ticketing app, or a vendor's dated admin console. Until now, automating around these meant either brittle custom scraping scripts, a costly system replacement, or accepting that someone on the team manually re-keys data between systems every day. AgentCore's Browser Tool gives a managed, sandboxed way for an AI agent to operate that old interface directly, essentially automating the human clicking-and-copying step without touching the underlying application. For an ops or support lead, this matters for a specific class of task: pulling status updates from a legacy tracking system into a CRM, filing renewals through an old vendor portal, or reconciling records across a system nobody wants to migrate. It doesn't replace a proper integration, but it closes the gap where integration isn't available or isn't worth building, and it's a capability worth flagging to whoever owns your process automation roadmap.

AWS has released the Amazon Bedrock AgentCore Browser Tool, a managed capability that lets AI agents interact with legacy web applications through the browser interface itself rather than through an API. The tool is part of AWS's broader AgentCore platform for building and deploying AI agents, and it's aimed squarely at a problem many companies have quietly worked around for years: business-critical software that has no integration path.

Legacy web applications, internal tools built a decade ago, vendor portals with no public API, older SaaS products that were never designed for programmatic access, are common in the operational stack of small and mid-sized B2B companies. When a business needs to move data in or out of these systems, the standard answer has been manual entry: someone logs in, clicks through the interface, and copies information by hand. According to AWS, the Browser Tool is designed to let an agent do that work instead, navigating pages, filling forms, and extracting information the same way a person would, but running as an automated, sandboxed process managed by AWS rather than a locally hosted browser automation script.

This differs from typical robotic process automation (RPA) tooling in one respect worth noting: the browser session is managed and isolated by AWS's infrastructure rather than run on a local machine or self-hosted server, which AWS positions as reducing the operational overhead of maintaining browser automation at scale. AWS's post walks through a reference architecture for building agents that authenticate into a legacy system, perform a defined task, and hand results back to a broader workflow, positioning the tool as one building block within Bedrock's AgentCore agent framework rather than a standalone RPA product.

For a company evaluating this, a few things are worth being precise about. This is not a way to add API-like functionality to a legacy app; it's a way to automate the manual browser interaction that a human would otherwise perform, which means it inherits the same fragility that any UI-based automation has: if the legacy application's interface changes, the automation can break. AWS's documentation frames the tool as suited to bounded, well-defined tasks, such as pulling a report, submitting a form, checking a status, rather than open-ended navigation of unfamiliar systems. Pricing and regional availability details are in AWS's documentation and were not fully detailed in the announcement; teams considering it should confirm current terms directly with AWS before committing.

The broader signal is that browser-level automation is now being folded into major cloud providers' agent platforms as first-party tooling, rather than left entirely to third-party RPA vendors or custom scripts. For companies with legacy systems in their stack, this lowers the technical barrier to trying agent-based automation on tools that were previously considered unautomatable without heavy custom engineering.