Your team has a Confluence page with a step-by-step process for an end-to-end approval workflow. It reads like a checklist that anyone could follow. But when you hand it to the RPA team, the bot halts on the first screen change, and the team is back in the rebuild cycle. The SOP sits untouched while manual work piles up. This is the maintenance treadmill that many enterprises see with legacy RPA: processes that should be automated are still run by people because the bots keep breaking.
Why RPA breaks here
RPA tools like UiPath, Automation Anywhere, and Power Automate automate by binding to selectors, XPaths, or object IDs. When an application updates its layout, the selector becomes invalid and the bot stops. IT teams report that an average of one in three bot incidents are due to a UI change. The cost is real. A mid-sized enterprise with 200 active bots might spend 30 to 40 percent of its RPA budget on rebuilds and maintenance rather than new automation. The team must identify the break, locate the correct new selector, rebuild the bot, test it, and redeploy. That cycle repeats every time a vendor releases a new release or the organization migrates to a different version of a software platform.
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. They move the mouse, click, type, read the result, and adapt when the UI shifts. That is why they survive UI and app updates, need no brittle selectors, and can recover from exceptions instead of halting.
From Confluence to running automation
A standard operating procedure written in plain English is already almost a prompt. A computer use agent can follow it directly, with no flowchart bot to build and babysit. You do not need a developer to map every element. You describe the goal and the steps, and the agent handles the execution. The agent can open the app, navigate to the correct screen, type data, click buttons, wait for results, and handle errors by reading the screen and adjusting its actions. It works across any application, including legacy systems and virtualized desktops where RPA struggles.
The durability advantage
Traditional RPA is brittle. When the app or UI changes, the bot breaks and a developer has to rebuild it. That is the maintenance treadmill. Computer use agents see the screen and act like a human. They move the mouse, click, type, read the result, and adapt when the UI shifts. That is why they survive UI and app updates, need no brittle selectors, and can recover from exceptions instead of halting. They also work across any application, including legacy systems and virtualized desktops where RPA struggles. Because they follow the SOP as written, you can iterate on the process without rewriting the bot each time.
How to move without the risk
You do not need to rip and replace everything at once. Start with one high-pain process that is still manual or frequently broken. Map the steps onto a Confluence page with clear objectives. Use a computer use agent to pilot the workflow. Measure the time savings, error reduction, and maintenance effort. Compare it against the existing RPA approach. When the pilot shows clear benefits, expand to other processes. RPA still fits very high volume, stable, deterministic, backend tasks. The win for computer use agents is the long tail, changing UIs, exception-heavy work, and SOP-driven processes. This phased approach lets you build confidence without betting the farm on a single technology.
SOPs are already prompts. The difference is how you execute them. Computer use agents see the screen, adapt to changes, and follow instructions as written. That makes them the durable way forward for the long tail of automation. To see how an agent can run your process from a Confluence doc, book a demo with the Coasty team at https://cal.com/coasty/15min .
Want to see this in action?
View Case Studies