Slack has announced Slack Code, a new product that brings AI coding agents directly into dedicated team channels, transforming what has been a solitary activity between one developer and one agent into a collaborative, visible process. The Salesforce-owned messaging platform is embedding agents from Anthropic, Cognition, GitHub, and Vercel into Slack channels where entire teams can observe, steer, review, and ship software together.

The shift represents a fundamental rethinking of how AI coding tools fit into enterprise workflows. Until now, most work with AI coding agents has happened invisibly, with individual developers working alone in terminal windows. Slack Code makes that work "multiplayer," allowing teams to see code diffs, live previews, and running plans in dedicated tabs, then archive the channel afterward with a searchable audit trail.

Why Is Slack Betting on Collaborative AI Coding?

The strategic argument underlying Slack Code is that artificial intelligence has inverted the economics of software development. Writing code used to be the scarce, expensive step. Now, according to Slack's executives, it is the cheap one, and the real constraint has moved upstream to human judgment.

"One of the things I love about this is that code is no longer the bottleneck. 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," said Rob Seaman, Slack's interim CEO.

Rob Seaman, Interim CEO at Slack

Cognition, the maker of the Devin coding agent, offered internal numbers to support this velocity claim. The company reported that its internal merged pull request count has gone up 10 times in recent months, while headcount has only increased by 40 percent. This suggests that AI agents are dramatically multiplying the output of existing engineering teams.

The pattern extends well beyond traditional software engineers. According to Cognition executives, many bugs are reported by sales teams directly in Slack, and technical team members then apply AI agents to fix them. This workflow bypasses traditional product management bottlenecks because the communication happens in real time, and the knowledge that fixes can happen quickly motivates faster reporting.

How Does Slack Code Actually Work in Practice?

When someone tags a coding agent from any conversation, the agent spins up a project-specific code channel and does the work in the open. The workflow resembles a newsroom more than traditional pair programming. In a live demonstration, Cognition showed how an agent could investigate a broken feature reported in an engineering channel, open a pull request, and even incorporate design files from Figma mid-task without breaking stride.

The agent finished by posting code changes alongside screenshots and a recorded demo proving the feature worked. This verification loop is central to Slack's pitch: cloud-based agents, unlike agents running on a developer's laptop, can generate an auditable record that the work is actually correct.

Steps to Deploy AI Agents in Slack

Slack is shipping a broader rework of how agents live in the product, making deployment simpler for enterprise teams:

• Agent DMs: Conversations with agents that behave like interactions with a colleague, allowing natural back-and-forth dialogue.
• Agents Tab: A new home base for every agent session with live status updates and a stop button for immediate control.
• Add to Slack Flow: Teams can deploy agents from platforms including Lovable, n8n, OpenAI, LangChain, and Airtable in a few clicks, with OAuth and configuration automated.

Slack Code is available on any Slack plan at launch, though customers need their own access to the partner agents.

What About Quality Control and "AI Slop"?

The obvious objection to democratizing software creation is quality. If anyone in a company can summon a coding agent, does an enterprise drown in what the industry calls "AI slop," meaning plausible-looking but poorly conceived output generated at scale by inexperienced users ?

Slack's executives argue that visibility is the antidote rather than the accelerant. When work happens in open channels, people can see intent, comment on work, and provide feedback in real time. Katie Steigman, Slack's VP of product, described her own practice as a template: when she puts pull requests up as a product person, she almost always tags in an engineer from her team, and that engineer provides specific technical guidance that the agent incorporates in a revision.

"The multiplayer part is a guard against that, actually, because people can see your work, people can comment on your work," explained Katie Steigman, VP of Product at Slack.

Katie Steigman, VP of Product at Slack

Rob Seaman framed the argument in broader terms: "I think the moral arc of multiplayer AI bends towards higher quality and less duplication." He pointed to Shopify, where CEO Tobi Lütke has written about restricting agentic coding to public channels precisely because it "immediately disseminates every single thing that's happening in the company" and levels the playing field.

However, skeptics have data on their side. Gartner predicted last year that more than 40 percent of agentic AI projects will be canceled by the end of 2027, citing escalating costs and unclear business value. This suggests that visibility and collaboration alone may not solve the fundamental challenge of demonstrating return on investment for AI coding tools.

Slack's bet is that the next chapter of the AI coding boom belongs to whoever owns the collaborative layer around those agents. By embedding agents directly into the communication platform where teams already work, Slack is positioning itself as the central nervous system for AI-assisted development, making the invisible visible and the solitary collaborative.