Skip to main content

Execution Modes

Recorded browser workflows can be used at different levels of maturity depending on how confident your team is in the process.

Early use: supervised execution​

When a workflow is new, teams often keep a person close to the run so they can confirm the result and catch anything unexpected.

This is a good fit when:

  • the workflow is still being tested
  • the portal changes often
  • the task has meaningful exceptions

Repeatable use: more independent runs​

As the workflow becomes more reliable, teams can reuse it more broadly for the same type of task.

This works best when:

  • the steps are stable
  • the inputs are clear
  • the team knows when a person should still review the result

What to monitor during execution​

When a browser workflow runs, teams usually watch for:

  • whether the task started in the right place
  • whether the workflow reached the expected end state
  • whether a person needs to review the output

Best practice​

Move from close supervision to broader reuse gradually. The goal is not to remove people too early. The goal is to use the Browser Agent where it is reliable and keep people involved where judgment is still needed.