At a major multinational, the shared services team spent a week rebuilding a bot each time the HR portal released a minor UI update. The same bot stopped at a missing checkbox or a changed field name and queued a ticket for a developer. The backlog of broken bots grew faster than the automation itself. In the end, the team paused the project and returned to spreadsheets and manual approvals.
Why RPA breaks here
Traditional RPA binds to selectors, xpaths, and object IDs. If the HR system adds a new column, reorders fields, or changes a button style, the bot breaks. Automation teams estimate that each selector change takes 2 to 4 hours of development time. For a global HR process with dozens of dependent bots, that adds up to weeks of unplanned work every quarter. The bot halts on an exception and waits for a human to decide what to do next. When the process depends on reading an error message that appears in a pop-up, the standard UiPath or Automation Anywhere solution must be rewritten to handle the new scenario, if it can be solved at all.
What changes with computer use agents
- Survives UI changes: The agent sees the screen, so it can recognize a button by its text or position and adjust its path without new selectors.
- No brittle selectors: You work with plain English prompts that describe what to do, not with fragile locator trees.
- Recovers from exceptions: When the agent encounters an unexpected state, it reasons about the current screen and tries a next step instead of failing.
- Follows the SOP as written: A standard operating procedure written in English is already almost a prompt. The agent follows it directly.
- Works on legacy and Citrix: Because the agent uses the screen, it can operate on virtual desktops, Citrix sessions, and older applications where traditional RPA struggles.
The one line a VP of automation should remember: Computer use agents survive UI changes and follow SOPs, so you stop rebuilding bots and start scaling automation across more processes.
How to move without the risk
You do not need to rip out all legacy RPA at once. Start with one high-pain HR process that has a clear SOP and frequent UI changes. Examples include benefits enrollment, expense approvals, or onboarding paperwork. Build a computer use agent to automate that workflow and compare the changeover cost to the current bot lifecycle. Measure the time saved, the reduction in failed runs, and the number of tickets the team no longer has to resolve. If the agent reduces failed runs by 70 percent and cuts the average fix time from days to hours, expand the approach to other HR processes. Keep the original RPA bots running for stable, high-volume tasks like payroll data imports, which still benefit from predictable selectors and deterministic flows. Over time, you can phase out the most brittle bots and rely more on agents that can adapt to changing systems. This phased approach lets you build confidence and real ROI before committing to a broader transformation.
If your HR shared services team is stuck rebuilding bots on every UI update, it is time to consider a more durable automation approach. Computer use agents can handle changing portals, exception-heavy workflows, and legacy systems without the constant maintenance treadmill. To see how a computer use agent works on your own HR processes, book a demo with the Coasty team at https://cal.com/coasty/15min.
Want to see this in action?
View Case Studies