Agents move into the channels where work happens
Every shift in enterprise software has moved the interface closer to the way people actually operate. Terminals preceded graphical user interfaces, which preceded the browser. The current shift puts language itself at the center: instead of clicking through configuration screens, users express intent and software responds.
That pattern is embodied in the AI agent. An agent doesn't wait for a user to navigate a UI; it interprets what someone wants and executes on it. For enterprises, the open question is no longer whether to adopt agents, but where those agents should operate. Vercel and Salesforce are answering that question by embedding agents directly into Slack, where collaboration already takes place.
"Vercel sets the bar for developer experience and agent infrastructure; Slack sets the bar for collaboration at work. This is a natural partnership as we work with more customers who want to deploy agents on Slack."
![]()
Rob Seaman, Chief Product Officer
Conversation becomes the application surface
Slack functions as the operating system for a large share of modern work. Decisions are made, projects are tracked, and escalations are raised within channels. Up to now, the plans hatched in those conversations were executed elsewhere, in separate systems, after the discussion concluded.
Agents change that sequence. An agent running on Vercel AI Cloud can live inside a Slack thread, listening, acting, and responding without interrupting the flow of discussion. If a team requests a report, an update, or a deployment, the agent fetches or triggers it on the spot. The conversation becomes a working interface rather than a preamble to one.
The runtime underneath
Vercel AI Cloud provides the infrastructure for these production agents: long-running execution, stateful sessions, and secured tool calls, combined with Vercel's existing speed and observability. Developers can build with Next.js, Hono, Nitro, or Express, using the AI SDK to wire up models, vector stores, and tools. Deployment handles message routing, context persistence, and security automatically.
Build an agent with Vercel and Slack
Start creating your own Slack Agent using the Bolt for JavaScript and Nitro template. Deploy it on Vercel to get a live, production-ready setup in minutes.
Deploy the template
v0 arrives in Slack
Vercel is also bringing its own agent, v0, into Slack as a demonstration of the model. v0 is described as a multiplayer, collaborative agent that works within the context of a Slack workspace while respecting permission boundaries. Teams can experiment, prototype, and ship sites, applications, and other agents from the same thread where they're already talking about them.
Be one of the first to use v0 in Slack
Try collaborative agent development in Slack. Build, test, and deploy together—powered by v0 and Vercel AI Cloud.
Signup for Early Access
Beyond Slack
The collaboration extends past Slack in the near term. Vercel and Salesforce plan to explore building with Agentforce 360 alongside Next.js and React over the coming months, with the goal of helping developers deploy more agents with less friction.
"The future of enterprise software is event-driven and conversational. Together with Salesforce, we're building the runtime for that future, where agents run on Vercel AI Cloud, interface through Slack, and connect directly to business logic in Salesforce."
![]()
Guillermo Rauch, CEO
The direction is consistent: enterprise applications have historically run in the cloud, but the next generation will run in the flow of conversation, turning collaboration itself into the interface for getting work done.



