Agents: what AI agents actually do | People in the Loop
The People in the Loop boardroom in daylight, seen from the doorway: a long table running away from you, an open book and a conference phone on it, walnut shelving on the left, a screen on the far wall, daylight through the window wall.
Everything in the room, tap to open At the table

Agents

Come in and have a look around

Four example AI agents are sitting at this table. Tap one to read what it actually does.

Nothing here is for sale. They are worked examples, written so you can see the shape of the job an agent is good at.

Looking for another room?

Examples

Agents most businesses need

Four examples, each one for a different industry. Tap one to see what work it would actually take on, and where a human stays in the loop.

Before anything gets built

Not everything needs an agent

Every real business process is three layers. Most of what teams call an AI problem is really not knowing which layer a given step belongs to. Sorting your work into these three is the step to do before a line of anything gets built.

So you are not paying agent-level cost for something a plain automation would do better, and not forcing a rigid automation onto a step that needs judgment. More on workflow vs. automation vs. agent →

Want to go deeper?

Everything we have written about working with AI agents is in the Library, free to read. Guides, a 76-term Lexicon, the comparison boards, and the blog.

Open the Library Nothing to sign up for, and nothing for sale.