Enterprise

Mainframe and Legacy App Automation with Computer Use Agents

Lisa Chen||6 min
+Z

You have hundreds of tasks sitting in a well-worn binder of standard operating procedures. They involve green-screen terminals, mainframe transactions, and outdated web portals. Your RPA bots can run many of them, but every time the vendor ships an upgrade or the internal team changes a field label, a developer has to rebuild the bot. Maintenance queues grow. Risk rises. And some processes never leave the binder because they feel too fragile for automation.

Why RPA breaks here

Traditional RPA, whether it is built on UiPath, Automation Anywhere, or Power Automate, relies on selectors, XPaths, and object IDs to know where to click and what to read. In a stable, modern SaaS environment, these can be brittle. In a mainframe or legacy app, they are often unreliable. Developers spend weeks on initial builds for processes that should take days. When a vendor updates the interface, the bot stops working and the team must rebuild it from scratch. A recent industry survey of automation teams reports that more than 60 percent of RPA maintenance time goes into selector updates and exception handling. The rebuild-on-change cost compounds quickly. Processes that should be repeatable and low-touch become high-touch projects. The same bots that once promised scalability now create a maintenance treadmill.

What changes with computer use agents

Computer use agents work differently. Coasty agents see the screen and act like a human: they move the mouse, click, type, and read the result. They do not need brittle selectors. If the UI changes, the agent adapts. If an exception occurs, it recovers instead of halting. This means they can run across any application, including legacy systems, Citrix desktops, and virtualized environments where RPA struggles. A computer use agent can follow your SOP directly. The SOP, written in plain English, is almost a prompt for the agent. The agent navigates the steps exactly as described, without the need for a separate flowchart bot or complex selector logic.

RPA needs stable, high-volume, behind-the-scenes tasks. Computer use agents are the durable answer for changing UIs, exception-heavy workflows, and SOP-driven processes on legacy systems.

How to move without the risk

You do not need to rip out all RPA at once. Start with one high-pain process that involves a legacy interface or frequent UI changes. Document the current SOP in plain English. Bring in a Coasty agent to run that process on a pilot environment. Compare maintenance time and failure rates against the existing bot. If the agent reduces downtime and developer effort, expand to similar processes. Keep your stable, high-volume backend tasks on traditional RPA where it still makes sense. Over time, the balance shifts toward agents for the long tail of work. This phased approach lets you modernize your automation strategy while keeping risk low.

Legacy systems do not need to hold back your automation program. Computer use agents give you a way to automate processes that are too fragile for RPA, with less maintenance and more resilience. Talk to the Coasty team to see how agents can handle your mainframe and legacy app workflows. Book a demo at https://cal.com/coasty/15min .

Want to see this in action?

View Case Studies
Try Coasty Free