Your team ships bots by the dozens, but the backlog never shrinks. Every time the procurement system updates its UI, three bots go down. Every time a form changes, developers spend days rebuilding selectors. Meanwhile, the SOPs that should be driving automation sit unread in shared drives. The result is a maintenance treadmill, not a scaling automation platform.
Why RPA breaks here
Traditional RPA tools like UiPath, Automation Anywhere, Blue Prism, and Power Automate rely on selectors, XPaths, and object IDs to locate elements on a screen. When a vendor ships a new release, the IDs often change, or the layout shifts enough to break the selector. According to industry surveys, around 30 percent of RPA development time is spent on maintenance. That means for every hour of new automation, three hours are spent fixing broken bots. A six month backlog is not a capacity problem. It is a structural problem with how bots are built and how they are maintained.
What changes with computer use agents
- Agents see the screen and act like a human: move the mouse, click, type, and read the result.
- They do not need brittle selectors or XPaths, so UI and app updates rarely break them.
- When the UI changes or a page loads slowly, agents recover instead of halting.
- A plain English SOP is already close to an agent prompt, so you can start from documentation without new flowcharts.
- They work across any application, including legacy systems, Citrix, and virtualized desktops where RPA struggles.
RPA is fine for high-volume, stable, backend tasks. Computer use agents are the durable layer for everything else: changing UIs, exception-heavy workflows, and SOP-driven processes.
How to move without the risk
Start with one high-pain process that mixes changing UIs and complex logic. Map it to an existing SOP, then run a pilot with computer use agents. Measure the time saved on maintenance versus the time spent on the pilot. If the net savings are positive, expand to a second process. Keep legacy RPA for the stable, high-volume cases, and layer agents on top for the rest. This phased approach lets you scale automation without rewriting everything at once.
The durability difference in practice
A finance team once automated invoice matching with RPA. Every time the ERP released a minor update, the bots failed and the team spent days restoring scripts. When they switched to computer use agents for the same process, the bots kept running through three ERP releases with almost no maintenance. The agents read the updated fields, clicked the new buttons, and completed the match without new selectors. The result was a drop in unplanned downtime and a backlog that finally started to shrink.
Why this matters for your center of excellence
Your center of excellence is judged not just by the number of bots, but by how many processes stay automated over time. When bots break every time a vendor updates a UI, you are not scaling automation. You are building a maintenance factory. Computer use agents let your team focus on new business problems instead of constantly rebuilding old ones. The difference is durable automation, not a treadmill.
If your RPA center of excellence has a six month backlog, the problem is structural, not capacity-related. Computer use agents provide a durable way to automate processes with changing UIs and SOP-driven workflows without the constant rebuild-on-change cycle. To see how agents can reduce maintenance and clear your backlog, book a demo with the Coasty team at https://cal.com/coasty/15min .
Want to see this in action?
View Case Studies