This page is machine-translated from French. Read the French original.

The agents

A agent is an authorized assistant to act, and not just to respond. Where an ordinary conversation produces text, an agent strings together concrete actions: searching the web, reading a file, querying a database, preparing an email, producing a graph.

Trigger an agent

In a conversation

Prefix your message with @agent :

@agent Cherche les dernières évolutions réglementaires sur la facturation
électronique en Suisse, puis résume-les en dix lignes.

The agent takes control until the task is completed, then returns control to normal conversation.

Permanently

A workspace set up in mode Automatic assigns each message to an agent. This mode is best suited for spaces dedicated to automation: it is slower and less predictable. → Conversations

Without human intervention

A planned task Executes an agent according to a schedule, without anyone being in front of the screen. → Planned tasks

How an agent works

1. He reads the request and determines what it must accomplish.
2. He chooses a tool among the skills activated in space.
3. He executes it and observes the result.
4. He starts again — as many times as necessary, each result guiding the next step.
5. He replies, based on what he actually achieved.

The sequence of events is visible in the conversation: you can see which tools were called and what they returned. An agent that fails will say so, rather than inventing a result.

A maximum number of tool calls limits loops: a blocked agent stops instead of running indefinitely.

What determines the quality of an agent

The model

This is the main factor. Using tools is more demanding than conversing: you have to choose the right tool, build valid arguments, and interpret a sometimes unexpected result. A model that answers questions very well can fail as an agent.

Each space can therefore assign an agent different model of the conversations. If agents matter to you, give them the most capable model you have available. → Choosing your models

The formulation of the request

An agent follows instructions. Vague instructions produce vague results.

Ineffective: «"Take care of competitive intelligence."»

Effective : «"Consult the Pricing pages of the following three websites, note the prices displayed for the entry-level offer, and present them in a comparative table with the date of consultation."»

Available skills

An agent can only do what its activated skills allow. A request that assumes a missing tool fails—and the agent must report it. → SKILLS

What an agent can do

Depending on the skills activated:

What to watch out for

An agent is truly taking action. A file writing skill actually writes files; an email skill prepares real messages. Activate only what is necessary, space by space.

Outgoing actions are regulated. Skills capable of sending something outward are designed to prepare rather than dispatch, so that a human can validate it beforehand.

An agent can be confidently wrong. It reports what its tools returned, without verifying it. An erroneous web page remains erroneous after being read by an agent.

An agent costs more than a response. Each step is a call to the model. With an external provider, a ten-step task costs approximately ten times that of a simple question.

One agent is slower. Allow anywhere from tens of seconds to several minutes depending on the task, model and equipment.

Expand the agents