# Putting It to Work: Automation and Teams

Asking for one task at a time is where everyone starts. The real value shows up when work runs on its own and several specialists share the load. This module is the plain picture of how that scales, and the words used for each step.

## From a task to a routine

When you ask for the same thing again and again, you can stop asking. Instead you set the work up once and let it run by itself, either on a schedule (every Monday, the first of the month) or when something happens (a new lead arrives, a payment fails). That setup is called automation, and the event that kicks it off is the trigger.

The shift is small but it changes everything: you move from doing the work to designing the work. You decide what should happen, when, and on what signal, and your AI employee handles it from then on. A good rule is that any task you describe more than twice is a candidate to turn into a routine.

- Schedule trigger: the work runs at set times, like a recurring reminder that actually does the job.
- Event trigger: the work runs the moment something happens, with no one watching the clock.

## Work that runs without you

An AI employee that keeps working between your check-ins, deciding its own next steps inside the boundaries you set, is acting as an autonomous agent. It is not waiting for the next message; it is making progress on the goal you gave it and bringing back results or questions when it needs you.

Autonomy is always bounded, and that is the point. You set the limits up front: what it is allowed to do on its own, what it must ask about first, and how much it can spend or send. Inside those limits it moves fast; at the edges it stops and checks with you. Good autonomy is less about handing over control and more about setting clear rules once, then trusting the work to run.

## Teams of agents

One AI employee can do a lot, but big jobs go better with a team. You can have several specialised agents, each focused on one kind of work, hand tasks to each other and combine their results. Coordinating them so the right one does each part, in the right order, is called orchestration, and the setup of many agents working together is called multi-agent. The helpers a lead agent delegates to are often called subagents.

It works like a small team with a lead. One agent breaks the goal into parts, passes each part to the specialist best suited for it, then pulls the pieces back into one finished result. You still talk to one point of contact; the coordination happens behind the scenes, the same way a manager runs a project without you meeting every person on it.

- Lead agent: takes the goal, splits it, and assembles the final result.
- Subagents: focused specialists that each own one part of the work and report back.

## Key takeaways

- Automation turns a task you keep repeating into work that runs by itself.
- A trigger is the signal that starts it: a schedule, or an event like a new lead.
- An autonomous agent keeps working between check-ins, inside the limits you set.
- Set the boundaries once: what it can do alone, what it must ask about first.
- Orchestration coordinates a multi-agent team, where a lead hands work to subagents and combines the results.

## Continue learning

- [Working With AI Employees](/en/free-ai-courses/working-with-ai-employees) — View the complete course.
- [Next: Trust, Quality and Responsible Use](/en/free-ai-courses/working-with-ai-employees/trust-quality-responsible-use)