Salesforce Launches Slack Code to Turn AI Development Into a Team Sport
The new feature integrates AI coding agents into shared channels, allowing cross-functional teams to review and approve code in real time.
Salesforce and Slack have launched Slack Code, a new feature designed to shift AI-assisted programming from a private activity into a collaborative team experience. The tool integrates agentic coding directly into the communication layer, allowing developers and non-technical staff to interact with AI agents in shared spaces.
At the center of the update are dedicated "code channels." In these spaces, teams can deploy integrated AI agents—including GitHub Copilot, Claude Code, Devin, and Vercel (v0)—to generate code and prototypes. To facilitate review, the system automatically produces "session artifacts," such as code diffs, Slack canvas documents, and live HTML prototypes. These artifacts allow team members to inspect and approve AI-generated work directly within the chat interface. According to Slack, the feature is available at no additional cost across all existing subscriptions.
Moving Beyond the Solo IDE
Traditionally, AI coding has functioned as a "solo player sport," with developers interacting with agents in private terminals or integrated development environments (IDEs). This isolation often creates a disconnect between high-level team discussions and the actual technical implementation.
Slack CMO Ryan Gavin noted that when AI work remains individual, organizations fail to get an "exponential return." By moving this process into the open, Salesforce is pursuing a broader "multi-player" strategy for workplace AI agents, a vision that also includes features like Claude Tag. Sateja Parulekar, VP of Product Marketing at Slack, stated that the company built code channels specifically to create a multi-player AI experience and bring more of the development lifecycle into the platform.
Reducing Lifecycle Churn
By integrating coding agents into the communication flow, Slack aims to reduce "lifecycle churn," the inefficiency that occurs when developers build features in isolation only to have them rejected during a final review. This shift effectively turns the development process into a cross-functional, "newsroom" style workflow.
This democratization allows product managers, designers, and marketers to initiate prototypes or provide real-time feedback without needing to leave the chat environment. To maintain technical integrity, Salesforce has included administrative guardrails. These controls ensure that non-technical users cannot ship code to production without a formal engineering review, preventing unauthorized or broken changes from reaching the live environment.
The Future of Agentic Workflows
As AI agents evolve from simple autocomplete tools into autonomous contributors, the focus is shifting toward how these agents are managed by human teams. Slack Code represents a bet that the future of software development is not just about faster coding, but about better alignment between technical and non-technical stakeholders.
Industry observers will now be watching to see how these integrated agents—particularly autonomous ones like Devin—handle complex, multi-file changes within a chat interface and whether the "multi-player" approach significantly reduces the time from prototype to production.