Skip to main content
Slack interface showing code embeds for Claude, Devin, Copilot, and Vercel AI agents in chat.

Editorial illustration for Slack Code embeds Claude, Devin, Copilot, and Vercel AI agents into chats

Slack Code Brings AI Agents Into Team Chats

4 min read

Slack announced a new product Tuesday called Slack Code that pulls AI coding agents out of solo terminal sessions and into shared channels where a whole team can watch the work happen. The Salesforce-owned company is launching with support for four agents: Anthropic's Claude Code, Cognition's Devin, GitHub Copilot, and Vercel's coding agent. Tag one from any Slack conversation, and it spins up a dedicated project channel, complete with code diffs, live previews, and a running task list, then archives itself once the job wraps, leaving a searchable record of what got built and why.

Slack Code is available on every Slack plan starting today, though users still need their own separate access to whichever partner agent they want to summon. The bigger idea is that AI coding right now is mostly a private exchange between one developer and one model, invisible to teammates, product managers, and designers who might have opinions about the result. Slack wants to make that process visible and collaborative instead, effectively turning agent-written code into something a whole channel can review and steer in real time, rather than a black box that shows up finished.

"One of the things I love about this is that code is no longer the bottleneck," Rob Seaman, Slack's interim CEO, said in a press briefing ahead of the launch. "Ideas, taste, judgment, craft — those are the things that are the bottleneck, and you've effectively extended the population that can contribute ideas, taste, judgment, and craft to anybody that exists in your Slack."

Why this matters

Slack is betting that AI coding works better as a team sport than a solo terminal session, and that's a genuinely different pitch from what Anthropic, Cognition, and GitHub have been selling. Bundling Claude Code, Devin, Copilot, and Vercel's agent into one channel means a founder can watch an agent write code while a PM flags a bug in real time, no context-switching required. For developers, the upside is visibility: code review becomes a group activity instead of a pull request nobody reads.

But the requirement that customers bring their own access to each partner agent is the catch worth flagging. Slack isn't replacing these tools' pricing or infrastructure, it's just building the room they meet in. That's a smart land-grab for Salesforce, since it makes Slack the default surface for AI-assisted engineering without owning the actual agents.

Worth watching whether teams actually shift workflows into chat, or whether this becomes another notification stream engineers mute within a week.

Common Questions Answered

What AI coding agents does Slack Code support at launch?

Slack Code launches with support for four AI coding agents: Anthropic's Claude Code, Cognition's Devin, GitHub Copilot, and Vercel's coding agent. Users can tag any of these agents from a Slack conversation to initiate collaborative coding sessions within shared channels.

How does Slack Code change the traditional coding workflow?

Slack Code moves AI coding agents from isolated terminal sessions into shared team channels where multiple people can observe and collaborate on the work in real time. This eliminates context-switching by allowing founders, PMs, and developers to review code, flag bugs, and provide feedback simultaneously within a dedicated project channel that includes code diffs, live previews, and task lists.

What does Rob Seaman mean by saying 'code is no longer the bottleneck'?

According to Slack's interim CEO, AI coding agents have removed code generation as the limiting factor in development, shifting the bottleneck to human judgment, creativity, taste, and craft. This allows more team members to contribute valuable ideas and feedback regardless of their coding expertise, effectively extending participation beyond traditional developers.

How does Slack Code's approach differ from existing AI coding tools?

While Anthropic, Cognition, and GitHub have primarily positioned their AI coding tools as solo developer assistants, Slack Code positions coding as a team sport by bundling multiple agents into collaborative channels. This enables real-time visibility and group code review instead of traditional pull request workflows, fundamentally changing how teams interact with AI-generated code.

LIVE03:17Slack Code embeds Claude, Devin, Copilot, and Vercel AI agents into chats