Slack AI integration that answers in the thread and acts across your tools
With Coworker's Slack AI integration, mention @Coworker in a thread and it answers there, using your Slack history and the other tools you connect, like Salesforce, Jira and Google Drive. A new message in a channel can start an agent, and agents post their results back to Slack as the Coworker bot or as you, with approval if you want it.
How Slack context reaches your AI tools
Coworker brings Slack together with the rest of your tools in OM2, its organizational memory, then answers in Coworker's own apps, or in Claude, ChatGPT and Cursor through Coworker MCP.
What teams do with the Slack AI integration
Coworker's Slack AI agents answer, file and post from the channels your team already uses.
Answers in the thread
Mention @Coworker with a question and get the answer in the same thread, run as the person who asked. Add +agent-name to send it to a specific agent.
Bug reports that file themselves
A new message in a channel you pick can start an agent, filtered by who sent it or what it says. A bug report in #support becomes a filed Jira issue.
Search Slack the way you see it
Search public and private channels, group DMs and DMs through your own Slack login, so answers only use conversations you can already see.
Results posted where people work
Scheduled and meeting-triggered agents post to a channel, a thread or a DM, like a Monday pipeline summary in #sales.
Channel admin on request
Ask Coworker to create a channel with a topic and invites, open a group DM, or add people to a channel.
More for team leads: Coworker for operations teams, and our guide to enterprise MCP.
Live in minutes
An admin adds Coworker to Slack
A workspace admin installs the Coworker app, so people can @mention Coworker and agents can post as the Coworker bot.
Each person connects their own Slack
A personal sign-in lets Coworker search the channels and DMs that person can see, and, where your workspace enables it, send messages as them.
Choose how agents send
For each agent, pick Send as Coworker bot or Me, limit where it can send, and tick Require approval to hold each message as a draft until someone approves it.
Every model needs the right context.
Slack's MCP server can only see what lives in Slack. OM2 connects those threads to the Salesforce accounts, Jira tickets, Gong calls, and Google Drive docs around them in one permission-aware knowledge graph, so Claude, ChatGPT, or Cursor starts every task with your customers, projects, and decisions already loaded.
Mosaic Financial
Series B fintech · $3.1M ARR · 47 seats
project
SOC 2 Type II audit
person
Priya Nambiar
workflow
Enterprise onboarding
team
Customer Success
source
#mosaic-impl
Mosaic Financial
Series B fintech · $3.1M ARR · 47 seats
Enterprise-grade security, from day one
SOC 2 Type II
Audited annually, report on request
Hosted locally
Your memory stays in your region. Open-source models run in the US.
Isolated, single-tenant
In our cloud or your VPC
Never trains on your data
No model provider, closed or open source
Your permissions, enforced
The access already set in Salesforce, Slack, Jira and Drive
Also included
Related Communication connectors
FAQ
Frequently asked questions
Coworker is a context layer for enterprise AI. It connects 50+ work tools, builds a permission-aware company graph called OM2, and runs agents that act across those tools, inside Coworker or in Claude, ChatGPT and Cursor through Coworker MCP.
Yes, through Coworker MCP. Coworker MCP gives Claude, ChatGPT and Cursor one connection to Slack and the other tools Coworker connects. Coworker's Slack search runs on each person's own Slack login, so it only returns conversations that person can already see. Slack also has its own MCP server, open to directory-published or internal apps.
Slack's official MCP server is a good fit if you only need Slack in Claude or Cursor. Coworker MCP is for answers that also need your CRM, Jira, Google Drive and call data: one connection to Slack and the rest of Coworker's 50+ connectors, with Slack search on each person's own Slack login.
Slack's own AI summarizes channels and threads, and its enterprise search covers Slack plus a set of connected apps. If most of your answers live in Slack, it is a good fit. Coworker is for questions that also need your CRM, ticketing and meeting data, and it can act on the answer, like updating a Salesforce record or filing a Jira issue, through 50+ connectors.
Message search across public and private channels, group DMs and DMs, channel history, and channel lookup. Searches run on the person's own Slack login, so they only return conversations that person can already see. Slack is also one of the sources Coworker's organizational memory draws on.
Send messages and thread replies as the Coworker bot or as you, edit messages you sent, create channels, open DMs and group DMs, add people to a channel, and update your own Slack status and profile. It does not delete messages, add reactions or schedule messages.
Yes. Mention @Coworker in a channel or thread. The question runs as you, with the tools you have connected, and the full answer posts back in Slack. Add +agent-name to send it to a specific agent.
Yes. Pick a channel, and optionally filter by sender or by text in the message. Each new matching message starts the agent, which runs as its owner and can post its result to a channel, a thread or a DM.
You decide. Turn on Require approval for an agent's Slack messages and each one is held as a draft until someone approves it. You also choose whether it sends as the Coworker bot or as you, and where it can send.
Yes. An agent can run on a schedule or when a meeting ends, then post to a channel, a thread or a DM. For example, a weekly list of deals past their close date, or open bugs for the release.
Coworker does this with a scheduled agent. Each morning, a Coworker agent can read Gmail, Salesforce and Slack, then send each account owner their open tasks and deadlines as a Slack DM, a channel post or an email. Messages can wait for approval before they go out.
Connect Slack in minutes
Connect your Slack data and watch agents go to work in minutes.