Salesforce brings Slack Code to all Slack plans as AI coding moves into shared channels
Salesforce’s August 24, 2026 Slack Code announcement is a product launch with 3 practical changes: AI coding agents now work inside shared Slack code channels, teams can review plans and diffs without leaving Slack, and human approval stays in the workflow before code reaches production. The news extends the Slack Code rollout covered over the weekend, but the strongest new detail is Salesforce’s confirmation that the feature is available across all Slack plans while partner agent access is licensed separately.
Salesforce Slack Code launches shared code channels for AI development
Salesforce’s August 24 announcement positions Slack Code as a new surface for collaborative AI software development. Instead of keeping agent work inside a private terminal, browser tab or isolated developer workflow, Slack Code creates a dedicated code channel for a specific task or project.
The channel is not just another chat thread. Salesforce says each code channel is organized around separate views for the conversation, the agent plan, line by line code diffs and a live preview of the running output. In practice, that changes the review pattern. Product managers, designers, developers and business stakeholders can see what the agent is attempting before the work is treated as done.
That is the real operational angle. AI coding agents have already made individual developers faster, but enterprise teams still need shared context, review trails and clear approval points. Slack Code is Salesforce’s attempt to make those controls native to the collaboration layer instead of adding them after the agent has already produced code.
Slack Code supports Claude Code, Devin, GitHub Copilot, ChatGPT and Vercel agents
Salesforce lists Anthropic’s Claude Code, Cognition’s Devin, GitHub Copilot, OpenAI’s ChatGPT and Vercel agents among the supported ecosystem for Slack Code. TechRepublic’s coverage of the Slack Code launch also framed the product as a way to pull AI coding agents out of isolated workspaces and into shared channels where teams can watch, guide and review work as it happens.
The commercial detail matters. Slack Code may be available across Slack plans, but integration still requires separate access to the partner coding agents. That means the Slack feature becomes the coordination layer, not a bundled replacement for agent subscriptions, repository controls, development environments or model usage policies.
For Salesforce, that is still strategically useful. Slack can sit above multiple AI coding tools without trying to become the only coding agent. The company is effectively making Slack the place where agentic coding work is requested, observed and approved, while the specialist agents continue to do the actual implementation work.
Slack Code adds human approval before AI generated code ships
The most important implementation safeguard is that Slack Code keeps people in the loop. Salesforce says channel members can pause, redirect or stop an agent during the task, and human signoff remains required before final changes are pushed to production.
That distinction is important because AI generated code creates a governance problem as much as a productivity opportunity. A developer can ask an agent to fix a bug quickly, but the team still needs to know which files changed, whether the agent followed the intended approach and who approved the work. InfoWorld’s analysis of Salesforce’s Slack Code strategy highlighted the same shift toward shared oversight, while also noting that adoption may depend on how developers respond to moving agent work into a more visible workflow.
That tension is realistic. Some developers will like the auditability and shared review pattern. Others may see it as more process around work they currently handle faster in their own tools. The practical adoption test will be whether Slack Code reduces handoffs without turning every small change into a crowded room.
Slack Code strengthens Salesforce’s agentic AI story through Slack
Slack Code gives Salesforce a more concrete answer to a question many enterprise buyers are asking: where should AI agents actually work? For coding, Salesforce is betting that the answer is not only the IDE or the terminal. It is also the collaboration space where requirements, feedback, approvals and business context already live.
TechRadar’s report on Slack Code described the feature as a way for teams to create focused channels for coding tasks, review diffs and preview output without leaving Slack. That is especially relevant for non developer stakeholders who may not be comfortable reviewing branches or local builds but can still validate whether a page, workflow or interface behaves as expected.
The immediate market implication is that Salesforce is using Slack to compete for the control layer around agentic work. GitHub, IDE vendors and model providers may own pieces of the development workflow, but Slack owns a large share of the conversation around work. Slack Code tries to connect those two worlds before agent generated work becomes too fragmented to govern cleanly.
Share With Others
Platinum
