A 40-step SOP for onboarding new contractors sat waiting for months at one large financial services firm. Every week, a junior analyst spent two hours walking through it. The process involved logging into three different systems, downloading files from legacy servers, updating spreadsheets, and emailing approvals. It was high value but too messy for the existing RPA team. Their bots were built on selectors and xpaths. When the HR portal refreshed its layout, the bot broke and the analyst had to resign themselves to manual work. This is the maintenance treadmill that stalls SOP automation and creates a quiet backlog of work that no developer ever touches.
Why RPA breaks here
Modern RPA platforms like UiPath, Automation Anywhere, and Blue Prism rely on selectors, xpaths, and object IDs to locate elements on a screen. These identifiers are brittle. When a UI refresh, a CSS change, or a minor layout tweak happens, the selector no longer matches and the bot halts. Industry surveys suggest that between 40 and 60 percent of RPA maintenance time is spent fixing broken bots after a change. In a complex 40-step process with multiple systems, a single UI change can break the entire workflow. Developers have to rebuild or retrain selectors, which delays work, increases cost, and discourages automation of anything beyond the most stable, backend processes.
What changes with computer use agents
- Agents see the screen and act like a human.
- No brittle selectors or xpaths required.
- They adapt when the UI changes without human intervention.
- They recover from exceptions instead of halting on the first error.
- They can follow an SOP written in plain English.
- They work across any application, including legacy and Citrix environments.
Selectors are brittle and rebuilds are expensive. Computer use agents see the screen and adapt, so you automate the SOP as written instead of fighting the UI.
How to move without the risk
You do not need to rip out all existing RPA at once. Start with a high-pain, exception-heavy process that is currently manual or partially automated. Document the SOP in plain language. Use a computer use agent to pilot it against a small, controlled set of cases. Measure the time saved, error rates, and maintenance effort compared with your current RPA bots. Once you see a clear win, expand to other processes that are too complex for traditional RPA. Keep legacy RPA for high-volume, stable, backend tasks where selectors are still reliable. This phased approach lets you prove the value of computer use agents without risking your entire automation program.
The durable path forward
The RPA market has matured. For high-volume, deterministic workloads, traditional bots still make sense. For everything else, changing UIs, exception-heavy workflows, and SOPs that live in natural language, computer use agents are the durable choice. They eliminate the rebuild-on-change cycle and let you automate processes that were previously out of reach.
See how a computer use agent can automate a 40-step SOP without writing a single selector. Talk to the Coasty team at https://cal.com/coasty/15min to book a demo.
Want to see this in action?
View Case Studies