Slack Turns AI Coding Into a Team Workflow With New Slack Code Channels

Slack Turns AI Coding Into a Team Workflow With New Slack Code Channels

Slack launched Slack Code, a new way for teams to work with AI coding agents inside dedicated Slack channels rather than limiting agent sessions to private terminals, browser tabs or standard threads. The feature is live today with integrations for Anthropic’s Claude, Cognition’s Devin, GitHub Copilot and Vercel, while OpenAI’s ChatGPT integration is coming soon.

Slack Code centers on a new workspace called a code channel. When a user mentions a supported coding agent on a project, Slack can create a channel specifically for that task and make the agent’s work visible to the wider team. People can follow its progress, review code changes, inspect previews and intervene while the task is underway. Once the project is finished, the code channel automatically archives. Its contents remain searchable, giving teams a record of the agent’s work and the conversations that shaped it.

Slack developed the format after encountering limitations with coding agents internally. Standard channel threads could become crowded during complex, multi-step tasks, while moving agent work into private tabs removed other team members from the process. Code channels are designed to provide more room for those sessions without separating them from the conversations already happening in Slack.

The company says more than 70% of its code channels are created and closed within one day, covering the process from an initial idea through a merged pull request. Slack is positioning the format for tasks including bug fixes and smaller interface changes that might otherwise enter a backlog.

The channels are designed to show more than an agent’s text responses. Code diffs and live previews can appear directly alongside the conversation, allowing teammates to examine work as it develops. A person can flag a problem, add a requirement or redirect the agent before it completes the task.

The feature also brings people outside engineering into the workflow. A product manager, designer or another teammate can contribute information to the same channel while an agent is working. In a demonstration involving Devin, for example, a designer added a Figma file during a task and the agent incorporated it into its work.

Cognition also demonstrated how Devin can respond when a problem is reported in an existing engineering conversation. The agent can investigate the issue, work toward a fix and open a pull request. In the demonstration, Devin identified the relevant code owner and brought that person into the process.

Jeff Wang, Cognition’s president of new enterprise, said the agent operates with its own computer and can use tools such as Chrome and DevTools to check its work. The completed task can include code changes, screenshots and a recorded demonstration. “Scaling things, auditing things, giving it to everybody — that is much easier with these cloud agents form factor than it is with local agents,” Wang said.

Each launch partner is using code channels somewhat differently. Anthropic’s Claude can move work out of a regular thread and into a dedicated channel, while summaries in the original conversation let teammates track progress. The channel can later be archived after the changes are completed.

GitHub’s integration is intended to let nontechnical teammates describe an issue in ordinary language, have an agent prepare a change and then bring an engineer into the channel for review. Vercel’s agent can generate a live preview after a change and place a shareable link into the conversation so teammates can examine the result.

Slack Code also uses Slack’s existing permissions and administrative controls rather than introducing a separate management layer. Changes that require additional oversight can be sent to a person for approval from within the channel.

Slack is making broader changes to how agents appear across its product as well. A new Agents & Tools tab provides a central location for agent interactions spanning threads, direct messages and code channels. The company is also working to make direct conversations with agents behave more like ordinary Slack DMs and to make their working status easier to follow.

The strategy reflects Slack’s view that AI-assisted software development should become a shared activity rather than a private interaction between one developer and an agent. Interim CEO Rob Seaman argued that faster code generation shifts more importance toward the decisions surrounding what gets built. “One of the things I love about this is that code is no longer the bottleneck,” Seaman said. “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.”

Cognition has seen a similar shift in its own development process. Wang said the company’s number of merged pull requests increased tenfold over the past several months while headcount grew about 40%. He described engineers assigning multiple tasks to Devin rather than waiting for one agent session to finish before beginning another.

Slack plans to take the code-channel concept beyond software development. The company says the same collaborative approach could eventually be applied to work such as marketing campaigns, contract reviews and IT onboarding.

For now, Slack Code starts with software teams and a group of established coding agents. Instead of replacing those agents, Slack is providing the collaborative environment around them, giving teams a shared place to assign work, supply context, review results and preserve a record of what happened.

This analysis is based on reporting from Slack.

Image courtesy of Slack.

This article was generated with AI assistance and reviewed for accuracy and quality.

Updated Aug 21, 2026

About this article: This article was generated with AI assistance and reviewed by our editorial team to ensure it follows our editorial standards for accuracy and independence. We maintain strict fact-checking protocols and cite all sources.

Word count: 908Reading time: 0 minutes

📧 Stay Updated

Get the latest AI news delivered to your inbox every morning.

AI News Daily

Breaking Intelligence • Since 2023

Join hundreds of thousands of AI professionals who start their day with our curated newsletter. Get breaking news, expert analysis, and exclusive insights.

Stay Ahead of AI

Get the latest AI breakthroughs, tools, and insights delivered to your inbox every week.

Free forever Unsubscribe anytime No spam guarantee

Go Premium

Unlock unlimited AI tools and an ad-free reading experience designed for AI professionals.

• Ad-free experience• Premium AI tools
Start Free Trial

14-day free trial • Cancel anytime
Plus $9/mo • Pro $90/yr (2 months free)

Follow Our Community

ChatAI

Breaking Intelligence

Your daily briefing on what matters in AI. Trusted by developers, researchers, executives, and AI enthusiasts worldwide.

© 2026 ChatAI. All rights reserved.