Trigger configuration

A trigger is the event inside a platform that starts a workflow. A Beam agent reads the triggers Summit exposes for a connected app, and when one fires, it picks up the payload, checks it against the conditions your team set, and starts the matching workflow. It records which trigger ran and what it carried. A trigger firing far more often than usual, or a payload missing fields the workflow needs, is paused and reported to a builder, so a person can look before automated steps run on bad input.

Action execution

An action is a write a workflow performs inside a platform, like creating or editing a record. When a workflow reaches an action step, a Beam agent reads the inputs, applies the rule your team defined, and runs the action against the target app through Summit. It logs the result so the run stays auditable. Actions that move money, delete data, or touch a system your team marked sensitive are held for a person to approve, since an irreversible write deserves a human check first.

Model record reads

A model is the structured data a platform exposes, such as an order or a contact. A Beam agent reads a model to gather the fields a workflow needs, joins them with data from other connected apps, and passes a clean set forward to the next step. It notes where each value came from for traceability. A model returning gaps, or records that conflict between two sources, is set aside for a person to reconcile, because a workflow built on wrong data spreads the error everywhere.