The agent comes back when it said it would.

If a customer asks it to "write to me on Monday", the agent schedules it on its own and picks the conversation back up on Monday. And whatever you leave scheduled goes out without anyone having to remember.

It only schedules what the customer asked for, quoting the phrase · 5 ways to repeat a task · Off by default, agent by agent

Follow-up depends on someone remembering.

It is not that the team does not want to do it. It is that the moment to write passed while they were handling something else.

It does not invent excuses to write again.

An agent that schedules its own follow-ups is an agent that pesters. So it can only schedule one when the customer asked for it, and it has to prove it.

Mondays at nine, without anyone remembering.

What the team wants to happen every week does not depend on any customer asking: it is set up in the panel, with its time and its time zone, and the agent runs it.

Everything scheduled is in plain sight.

The tasks of an agent live on their own screen: Virtual Office → the agent → Tools → Scheduled tasks. Four columns and, inside each task, what happened on each run.

An agent that writes on its own needs brakes.

The feature exists so no customer is left waiting for a message. Everything else is there so it does not turn into something else.

What gets checked in the first demo.

The module that comes next.

Bring a customer who was left waiting.

In the demo we schedule the follow-up on that conversation and set up the task your team currently does by hand every Monday.