Your operations team writes standard operating procedures in plain English. A developer then builds a bot on top of them using selectors, xpaths, and object IDs. When the application updates, the bot breaks and a developer must rebuild it from scratch. That is the maintenance treadmill. Many enterprises report that over 30 percent of their automation budget goes to rework and fix broken bots rather than new work. Meanwhile, the SOP sits unchanged, creating a gap between what is written and what runs.
Why RPA breaks here
Legacy RPA platforms like UiPath, Automation Anywhere, and Blue Prism rely on brittle selectors and object IDs. When a UI changes even slightly, those selectors fail and the bot halts. You see this in three common patterns. First, UI updates force a rebuild of every affected step. A single release can break dozens of bots across different applications. Second, selector conflicts arise when multiple elements share similar attributes, causing random failures. Third, RPA struggles on legacy systems, Citrix virtual desktops, and terminal-based apps where selectors simply do not exist. Industry studies on RPA maintenance show that unplanned maintenance can consume up to 40 percent of a program’s total cost of ownership. Your automation backlog grows not because you lack ideas, but because the underlying approach cannot keep up with change.
What changes with computer use agents
- Survives UI changes
- No brittle selectors
- Recovers from exceptions
- Follows the SOP as written
- Works on legacy and Citrix
Computer use agents see the screen and act like a human, so they stay aligned with your SOPs even when the application changes.
How to move without the risk
You do not have to rip and replace all your automation overnight. A pragmatic path works well. Start by identifying one high‑pain process where the SOP is written in plain English and the UI changes frequently. Examples include onboarding workflows, exception handling for returns, or data reconciliation across multiple systems. Run a pilot with a computer use agent to see how it handles those steps. Measure not just speed but also stability and how often the agent needs human intervention. If the pilot succeeds, expand to related processes. This phased approach lets you keep your existing RPA for high‑volume, stable, deterministic tasks while computer use agents take over the changing, exception‑heavy work. Over time, you can migrate more work to agents, freeing your RPA developers to focus on new initiatives.
The durable way forward is to keep your SOPs and automation in sync. Computer use agents can follow the doc directly, adapting to every UI change instead of breaking. Ready to see how your own processes would run with an agent that follows your SOPs? Book a demo with the Coasty team at https://cal.com/coasty/15min .
Want to see this in action?
View Case Studies