Saved process replay

Twin captures a task once so it can run again without repeating the manual steps. A Beam agent starts a saved Twin process on the schedule or event your team set, reads the inputs it needs from the connected source, and lets the run proceed under the rule you approved. Runs that finish cleanly are recorded and passed downstream. A run that returns an error, produces output outside the range you defined, or hits an input it was not built for is stopped, and the agent routes the case to a person for a decision.

Run output validation

Twin produces output each time a saved process runs. A Beam agent reads that output, compares it against the checks your team defined, and forwards it to the next system where permissions allow, for example handing a completed result to your CRM. Output that passes moves on automatically. Anything that fails a validation, comes back empty, or differs sharply from the previous run is held. The agent notifies the owner with the run details attached, and a person decides whether to accept the result, rerun the process, or fix the input.

Recurring run starts

Twin exists to remove work you would otherwise repeat by hand. A Beam agent listens for the conditions your team tied to a Twin process, a new file, a due date, an upstream update, reads the relevant data, and begins the run where your account allows. Well-defined, recurring cases fire on their own. A trigger that arrives with missing context, matches more than one process, or falls outside the approved window is not run automatically. The agent surfaces it to a person, who confirms which process should handle it.