Triggers

A Loop needs at least one trigger. You can add more than one, and each trigger can start a separate run.

Manual trigger

Use Manual trigger for work that should only start when a person chooses Run. A manual trigger has no event payload.

This suits one-off or deliberately supervised work, such as searching a CRM segment and preparing drafts for a review checkpoint.

Schedule

Use Schedule for an active Loop that should start at a recurring UTC time. The friendly picker supports daily, weekdays, and weekly schedules. Advanced mode accepts a five-field cron expression.

Common examples:

• 0 2 * * * runs daily at 02:00 UTC.
• 0 9 * * 1 runs Mondays at 09:00 UTC.
• */15 * * * * requests every fifteen minutes.

The dispatcher checks a five-minute window. Do not schedule a cadence tighter than five minutes, and allow for a short enqueue delay after the exact minute.

Pausing the Loop stops new scheduled runs. It does not cancel a run that has already started.

Event

Use Event for an active Loop that should start when Fondaro records one of these CRM events:

• lead created;
• lead status changed;
• lead assigned;
• deal created;
• deal stage changed;
• deal won or lost;
• task created or completed;
• note created.

The event payload contains identifiers for the changed record. For lead events, connect the trigger to Lead from event before an AI node that requires a full lead.

Repeated deliveries are deduplicated in a five-minute bucket. Event starts are also capped at 25 runs from one source event and 100 event runs per organization per rolling minute. This protects the account during a large import, but it means event Loops are not an unbounded bulk-processing mechanism.

Pausing the Loop stops new event runs. It does not cancel runs already in progress.

Test a trigger

Open the Loop editor and choose Test this loop or the run-options menu.

• A manual-only Loop starts immediately.
• For a schedule, choose Run now to start the scheduled path without waiting for its next UTC occurrence.
• For a lead event, search for a real lead to test the downstream path with that lead's ID.
• Test with a synthetic lead only exercises trigger routing. It will fail when the graph reaches Lead from event or another node that needs a real CRM lead, so use a real lead for an end-to-end rehearsal.

Testing creates an ordinary run record. Follow it on the canvas and in the recent-runs strip.