Salesforce Launches Slack Code to Move AI Development Into Shared Channels

Salesforce has released Slack Code, a new service that integrates AI coding agents directly into shared Slack channels. The move aims to transform AI development from a private, single-developer experience into a collaborative, cross-functional process.

Slack Code functions by automatically creating project-specific "code channels" whenever a supported coding agent is tagged in a conversation. These channels serve as a centralized hub where developers, product managers, and designers can interact with the AI in real-time. At launch, the service supports agents from Anthropic (Claude), Cognition (Devin), GitHub Copilot, and Vercel, while an integration for OpenAI's ChatGPT is listed as available soon.

To facilitate oversight, the shared workspace provides specialized views that allow stakeholders to review the agent's progress without needing to enter a code editor. These include dedicated views for the agent's plan, code diffs, and live previews—such as HTML previews provided via Vercel. Once a task is finished, the agent can package the work for sign-off or open a pull request, at which point the channel is automatically archived to serve as a searchable record and audit log.

Collapsing the Development Pipeline

Traditionally, software development follows a fragmented pipeline: a request begins in a collaboration tool like Slack, moves to a project management system like Jira, and is finally executed within a developer's integrated development environment (IDE). This "one-person, one-agent" model often creates silos and communication gaps.

By moving the agent into the collaboration layer, Salesforce is attempting to collapse this pipeline. This shift allows non-technical stakeholders to initiate and monitor development work directly, potentially reducing the friction of handoffs. Manoj Chandra Jha, a principal analyst at Nord-IQ Research, notes that this model could lead to faster fixes, improved cross-functional input, and fewer costly handoffs between technical and non-technical staff.

The Trade-off Between Speed and Focus

While the integration promises efficiency, it introduces a tension between collaboration and the nature of software engineering. The primary benefit is the ability to catch incorrect assumptions early through increased visibility across different roles in the product cycle.

However, critics argue that the environment may clash with the requirements of technical work. Advait Patel, a senior site reliability engineer at Broadcom, points out that coding is "deep work," whereas Slack is often viewed as an "interruption machine," suggesting that merging the two on a single surface is not a guaranteed win. Additionally, the proliferation of task-specific channels could lead to notification overload and potential security risks if repository access is conflated with channel membership.

What to Watch

As teams begin adopting Slack Code, the industry will be watching to see if the gains in velocity outweigh the potential for developer distraction. The upcoming arrival of ChatGPT integration will further expand the ecosystem of available agents, potentially increasing the tool's adoption across diverse technical stacks.