You invest in IDP to read documents and extract data, and RPA to move that data through systems. But in practice, the bots break every time the UI changes, and the backlog of fixes grows. The real value is in the process, the SOP on a spreadsheet or wiki, but RPA can't follow it. A computer use agent can.
Why RPA breaks here
RPA bots rely on selectors, XPath, and object IDs. When a vendor refreshes a portal or changes a field label, the bot stops working. Most enterprises see at least 15 percent of bot tickets tied to UI changes each year. That means a developer has to rebuild the automation, which costs time and introduces new bugs. In exception-heavy workflows, the bot halts on the first unknown state, and a human must step in. The SOP is still the source of truth, but it can’t run the process.
What changes with computer use agents
- Survives UI changes. Agents see the screen and act like a human, so they keep working when selectors no longer match.
- No brittle selectors. You don’t need to maintain object models for every screen.
- Recovers from exceptions. If a page looks different or an error appears, the agent can read it and choose a next step instead of stopping.
- Follows the SOP as written. The process described in plain English is already almost a prompt. The agent can read it, interpret the steps, and execute them.
- Works on legacy and Citrix. Agents run on real desktops and can operate in environments where RPA struggles, including virtualized desktops and thin clients.
RPA is still best for high-volume, stable, backend tasks. Computer use agents are the durable way forward for anything that changes, needs exception handling, or is defined by a human-written SOP.
How to move without the risk
You don’t have to rip out all your IDP-plus-RPA investments at once. Start with one high-pain process where the UI changes often or the process is documented in an SOP. Run a pilot of the computer use agent on that process. Measure the difference in uptime, fix tickets, and time spent by people. Once you see the pattern, expand to similar workflows. Keep your proven RPA for the stable, high-volume tasks, and let agents handle the rest. This phased approach lets you build confidence without betting the whole operation on a new approach.
Computer use agents can follow the same SOPs you already have, without the need for brittle selectors or constant rebuilds. Book a demo with the Coasty team to see how a single agent can handle your changing document workflows. https://cal.com/coasty/15min
Want to see this in action?
View Case Studies