Slack launches collaborative AI coding tools to capture enterprise software market
Salesforce’s Slack has launched a product embedding AI coding agents directly into team channels, betting that collaborative visibility will solve the enterprise sector's struggle to prove return on investment for artificial intelligence.
Salesforce-owned Slack announced Slack Code, a new product that embeds artificial intelligence coding agents directly into dedicated team channels. The launch integrates tools from Anthropic, Cognition, GitHub, and Vercel, shifting software development from isolated terminal windows into shared, visible workspaces.
When a user tags an agent, the system creates a project-specific channel where the software writes code, generates live previews, and maintains a running plan. The channel archives upon completion, leaving a searchable audit trail while ensuring the agent inherits the invoking user's exact security permissions without gaining elevated access.
Slack interim chief executive Rob Seaman argued that this shift changes the fundamental economics of software creation. He noted that writing code is no longer the bottleneck, positioning human judgment and ideas as the new constraints that the platform aims to democratize across an entire organization.
Cognition president Jeff Wang provided internal metrics to illustrate this velocity, stating that merged pull requests increased tenfold over recent months while headcount grew by only forty percent. He predicted that routine engineering tasks will be automated away, leaving humans to focus on business logic and creative planning.
Addressing concerns about low-quality AI output, Slack vice president of product Katie Steigman argued that public visibility acts as a natural guardrail. She explained that team members can review and correct AI-generated work in real time, a practice Seaman noted is already championed by Shopify chief executive Tobi Lütke.
The launch arrives as enterprise buyers remain highly skeptical about the financial returns of artificial intelligence. A recent McKinsey survey found that while sixty-two percent of organizations are experimenting with AI agents, only thirty-nine percent report any actual bottom-line impact.
Furthermore, Gartner forecasts that more than forty percent of agentic AI projects will be canceled by the end of 2027 due to escalating costs and unclear business value. Slack is attempting to bridge this gap by selling the collaborative infrastructure around these agents, rather than just the underlying models.
Beyond dedicated code channels, the update introduces agent direct messages that mimic colleague conversations and a centralized tab for monitoring agent status. The company also streamlined the deployment process, allowing teams to integrate agents from platforms like OpenAI and LangChain with automated configuration.
