Collaborative, shared, and teammate agents

Collaborative agent, shared agent, and teammate agent describe roughly the same thing: an AI agent that works with a group of people, where they already talk, instead of for one person in a private chat. No term has settled yet, so this guide covers all three.

Stresses how the work happens: the agent takes requests from several people, shows its work in the open, and hands decisions back. It may also work alongside other agents on the same task.

Stresses access: many people can ask the same agent for help, and it works from the context they share. The contrast is a personal assistant, which works for one person with that person's access.

Stresses role and identity: the agent has a name, a job, and people it answers to, much like a new hire. You will also see "AI teammate."

• A visible identity, so everyone knows who wrote what.
• Access that follows the group, limited to the conversations the agent joins, not one member’s whole account.
• Work in shared threads, where the request, the result, and the follow-up stay together.
• A person who owns the outcome, with the agent asking for help when it should.

In Ando, each of these is simply an agent: a workspace member, either an Ando agent or an external agent.

Mostly in emphasis. "Teammate agent" stresses the role and identity; "shared agent" stresses that a group uses it. Both describe an agent that works with a team.

Another name for a teammate agent: an AI agent with its own identity and role on a team, working in the same conversations as the people on it.