(Co)agents
collaborative coworkers

Give every conversation a few more contributors.

(Co)agents™ are AI teammates that contribute to group discussions the way good colleagues would. They look things up, bring in what's missing, and offer ideas and reasoning nobody has yet raised. They inform. They don't decide.

Most group decisions are made with whatever knowledge happens to be in the room.

A team debates an important issue for an hour. Critical information exists in the corporate database, public filings, or online sources. By the time someone goes and checks, the discussion has moved on and the team needs to backtrack. The limitation was never the people. It was that people can't talk and search at the same time. (Co)agents can.

OBSERVE group discussion in real time TRACK human insights and reasoning IDENTIFY critical information gaps SEARCH for responsive intelligence EXPRESS conversationally to group (Co)agents are active listeners, searchers, and contributors in real time.
Swipe to see the full cycle

(Co)agents close that gap from inside the conversation, while it is still happening.

What They Are

Coworkers, not chatbots.

A chatbot sits behind a window, waiting for you to ask it something. (Co)agents sit in the room with the team, following the discussion as it unfolds and participating in it directly. The distinction is not cosmetic. Invented by Unanimous AI, (Co)agents observe the group conversation, find gaps in knowledge or reasoning, and jump in with helpful insights.

(Co)agent

An AI contributing coworker that joins a groupwise discussion to supply knowledge, retrieve information, or generate ideas, without taking the decision away from the people in the room.

Three Types

Three kinds of contribution.

(Co)agents differ by where their contribution comes from. A team conversation can include none, one, or all three, and the mix can change from session to session.

Knowledge

Knowledge (Co)agents

Draw on your organization's own material: the docs, records, and data your team already relies on. When information exists inside the company, it arrives in the conversation instead of in someone's follow-up email.

Scouting

Scouting (Co)agents

Search beyond the organization for current information from the open internet. Published research, market figures, regulatory changes, and competitor activity, retrieved while the discussion is still live.

Brainstorming

Brainstorming (Co)agents

Generate original contributions rather than retrieving them. They widen the option set, introduce alternatives the group has not considered, and keep an ideation session from narrowing too early.

Two Modes

Ask them in, or let them speak up.

(Co)agents can wait to be asked or proactively jump in. You set the mode.

Reactive

Waits to be asked. A team member calls on the agent directly and it responds in the thread, the way you would ask the one person on the team who already knows. Predictable, unobtrusive, and fully under the group's control.

Proactive

Watches the discussion and interjects when it recognizes an opportunity: a factual question left hanging, an assumption worth grounding, a brainstorm that has started to circle. It offers what the moment calls for, then gets out of the way.

Three agent types across two interaction modes give your team exactly what they need.

Type Reactive Proactive
Knowledge Someone asks what the current policy actually says. The agent pulls it and answers in the thread. The group starts debating a question the organization has already settled. The agent surfaces the existing answer before the debate goes further.
Scouting Someone asks what a competitor announced this morning. The agent goes and finds out. A claim about market size passes unchallenged. The agent brings in the current published figure.
Brainstorming A subgroup runs dry. Someone asks for more options and gets a fresh set to react to. Ideas begin converging early. The agent introduces an alternative nobody raised.
Where They Work

They join the discussion your team is already having.

(Co)agents are not tied to a single platform or to a particular size of group. They contribute to an eight-person channel the same way they contribute to a deliberation of several hundred.

Thinkscape®

Built in. (Co)agents participate across parallel ThinkTanks in a full Thinkscape deliberation.

Microsoft Teams (beta)

Deployed into the channels and meetings your organization already runs.

Slack (beta)

Present in the working channels where much of the real discussion happens.

By API

Embedded in your own systems and internal tools alongside the rest of the Unanimous AI stack.

Humans In The Loop

They contribute. Your team decides.

(Co)agents have no vote and no veto. They cannot commit the group to anything. Their job is to ensure human teams are well-informed by providing efficient contributions that are on-task and on-target. They support teams without replacing any human voices in the decision-making loop.

Visible, not hidden

Contributions appear in the conversation, attributed and open to challenge like anyone else's.

Additive, not authoritative

An agent widens what the group is considering. It never narrows it on the group's behalf.

Configured by you

Which types run, in which mode, and with access to what, is set by the organization.

Trusted By Mission-Critical Teams At

Add a contributor to your next conversation.

See (Co)agents working inside a live team discussion.

Get a Demo