Slack Code: AI Coding Agents Get Shared Channels for Review and Oversight

Slack has introduced a new feature called Slack Code, which introduces AI coding agents to dedicated project channels for review and oversight. By doing so, Slack aims to make it easier to track and manage the work of AI coding agents, which can write software surprisingly quickly. Previously, it was more challenging to keep tabs on what these agents were doing as they operated in isolated terminals and browser tabs.

Slack Code will make these changes visible to the entire team, allowing developers and other team members to watch, guide, and review the AI agents' work in real-time. Currently, the feature is available across all Slack plans, but users need access to the specific AI agents they wish to use. Initially, Slack Code will focus on software development, with integrations involving AI agents from Anthropic’s Claude Code, Cognition’s Devin, GitHub Copilot, and Vercel.

OpenAI’s ChatGPT integration is expected to be added soon. The company is also expanding the experience to users who typically do not write code, such as non-technical employees. By having an AI agent work on a fix and notify a human engineer to review the changes, Slack aims to reduce the time it takes to spot and resolve issues.

Slack believes that increased visibility into AI-generated code can help catch errors and quality issues before they become production problems. However, the visibility alone does not guarantee the quality of the resulting software, and human oversight is still necessary. Slack has implemented existing user permissions and administrative controls for AI agents to ensure safe access to high-stakes changes.

As the company rolls out Slack Code, it also introduces an "Agents" tab to manage agent conversations, agent direct messages (DMs), and tools to integrate third-party agents. This suggests that Slack is positioning itself as a shared workspace where humans and AI can work together, rather than just a place for discussing AI-generated work.

For developers, Slack Code changes the environment for AI-assisted coding by integrating agents into shared channels where teammates can follow changes, review decisions, and step in when needed. While this may make collaboration easier, developers should still treat AI-generated code with the same care as any other contribution by reviewing, testing, and understanding the changes before it reaches production.

The real value of Slack Code may lie in making AI agents more manageable and easier for developers to inspect, challenge, and improve.