Slack · Teams

A teammate in the channel.

Not another AI tab that everyone briefs alone. Tag 1Presence in a channel and it works in the open — carrying what that room is about, answering from the company's memory, on the company's account.

Everyone has their own AI tab. Your team has a teammate.

The cost of an assistant that lives in a private tab is invisible until you count it: every person briefs it separately, every answer is written for an audience of one, and nothing anyone learns is available to anyone else. Work that belongs to a team ends up rebuilt, quietly, several times over.

Putting 1Presence in the room inverts that. It reads what the channel is about, does the work where everyone can watch, and leaves the answer where the conversation already is. The next person picks up from what it just said instead of starting again.

This is not the Slack connector

Both are called Slack and they do genuinely different jobs, so it is worth being exact about which one you want.

The Slack connectorThe team agent
Whose it isYours. It acts as you.The company's. It acts as itself, in the room.
Where it reachesEvery workspace you belong to — client, community, company.The channels you invite it to, in one workspace.
Who can use itOnly you. It works inside your own 1Presence.Anyone on your team who is in the channel — and everyone there sees the answer.
What it knowsYour memory, your vault, your connected tools.The company's memory, plus whatever that channel has built up.
Who paysYour own plan.The organisation's balance.

Most teams end up with both: the connector so each person's own 1Presence can read their workspaces, and the team agent so the company has a presence where the work is actually discussed.

What it actually does

  • Tag it, and it gets to work in the open. It acknowledges in seconds, then edits that same message into a running checklist as it goes — so the channel can see what it is doing rather than watching a silence. The answer lands in the thread when it is done.
  • The room is the context. Each channel has its own memory. It knows what this channel is about without anyone re-explaining, and a second person can follow up on the first person's request without briefing it from scratch.
  • A different channel is a different memory. What it learns in one room is not carried into another, and none of it touches anyone's personal 1Presence. Rooms, with their own keys.
  • Each channel can have its own specialist. Assign an agent per channel and a bare tag runs that one — the researcher in the research channel, the support drafter where customer threads live. Name a different agent inline when you want someone else on it.
  • A direct message is still the company's Presence. Message it privately and you get the org's agent, reaching the org's knowledge on the org's account — a private channel of the company, not a personal session. That conversation gets its own memory too.
Tagged in the room, answering in the room — and the follow-up nobody had to re-brief.

How to address it

Invite it to a channel the way you would any Slack app, then tag it with what you want — in plain words, the way you would ask a colleague. A bare tag runs the agent assigned to that channel. Adding an agent's name after the tag runs that one instead.

One practical detail worth knowing before you type: Slack decides the exact handle when the app is installed, and it can differ between workspaces — if something in your workspace already took the obvious name, Slack hands out a variation. So 1Presence reads the real handle back from Slack at install and shows it to you in the admin panel, next to the invite instruction and the channel list. Copy it from there rather than guessing, and if your workspace ended up with a look-alike, the panel says so plainly and names the one that can actually hear you.

Who it thinks it is talking to

An admin maps each person's Slack user to their 1Presence membership once. After that the agent knows who is asking, and everything that follows from identity applies: it addresses them by name, it stays inside the teams that person belongs to, and their usage is attributed to them rather than to an anonymous room.

Colleagues you have not mapped yet can still use it — they are simply not recognised individually, and their work is attributed to the organisation rather than to them. Mapping is worth doing for the regulars and skippable for everyone else.

People from outside your workspace cannot. Guests and outside collaborators in a shared channel are turned away with a short message, because the team agent spends the organisation's credit and reaches the organisation's knowledge — that is for the team, not for whoever happens to be in the room.

What an admin controls

Which agent answers where

Per channel, choose the agent a bare tag runs. People can still name another agent directly, and the list only offers agents your organisation actually has.

What it is allowed to reach

An agent brought into a channel starts with no access to your tools. The first time it needs one it says so in the channel and waits — permission is granted on the agent's Access tab in 1Presence, deliberately, not in the heat of a thread.

What it can spend

Work in Slack draws on the organisation's balance. When that balance is out, runs stop and the channel is told plainly — no silent failures, and no surprise overage.

What it has been doing

A running list of every run: who asked, in which channel, which agent, what it reached for, how it ended. The record deliberately keeps no message text — it tells you what happened, not what was said.

Two of them sit on the bot's own tile in your organisation's admin panel, under Connectors: the channel assignments and the activity list. The other two live where they belong rather than duplicated here — an agent's access is granted on that agent, and the balance is your organisation's ordinary billing.

It answers; it never starts

Worth stating precisely, because it is the thing teams are right to ask about. Every message it sends is a reply to someone. It has no standing watch on your channels, sends no unprompted pings, and posts nothing on a schedule.

It replies in two situations. The obvious one is when you tag it. The other is a short courtesy: for a little while after it has answered — in the thread it replied in, or to the same person who just asked it something in that channel — you can carry on without tagging it again, the way you would with a colleague who is clearly still listening. That window is measured in minutes and then closes, and it applies in public channels; in a private channel, tag it each time.

Everything else your team says is discarded the moment it arrives — not stored, not read by an agent, not acted on. Slack delivers channel activity to any app that has been invited into a channel; the difference is what the app does with it, and outside those two windows the answer here is nothing at all.

Where the work is discussed

One shared brain, in the shared room.

The team stops briefing five separate assistants and starts building on one. It carries each channel's context, answers where everyone can see it, and does it on the company's account, under the company's rules.

1
memory per channel, kept apart
0
unprompted messages, by design

Start a conversation.

Free to try. No credit card. Just you and your agent.

Works on any device. Takes 60 seconds to start.