← Blog

Vercel Board Is a Coordination Layer for Agent Teams, and Designers Should Pay Attention

Vercel just shipped Board, a shared message board that lets multiple AI agents coordinate work, claim tasks, post updates, and hand off results without stepping on each other. Here is why that matters if you are building with AI today.

By ProbG · November 20, 2026

Vercel just shipped Board, a purpose-built coordination layer that lets teams of AI agents share tasks, claim work, post progress, and hand off results, all in one place. It is experimental, but it is already running in production at Vercel, and it signals something real about where AI-assisted building is heading.

The thesis here is simple: until now, one agent doing one thing at a time was fine. Board is evidence that the next phase involves fleets of agents working in parallel, and the designers who understand that shift early will build things that feel noticeably smarter.

What Board Actually Is

Think of it like a Slack workspace, but for agents rather than people. It has channels, threads, and labels. An agent picks up a task, claims it, posts updates as it works, then marks it done and shares the result. Another agent, or a human, can then review and pick up where it left off.

Vercel built it to solve a coordination problem they ran into internally. A single agent can keep notes in a text file, but there was no shared, structured place for many agents to work together without colliding. Board is that place. It is "headless" meaning it has no fixed visual interface baked in; you connect it to whatever tools your team already uses.

Two real use cases Vercel shipped with this: a swarm of security agents that scan codebases for vulnerabilities and push fixes, tracked entirely through Board; and v0's "software factory" which Vercel says now merges more than 50% of pull requests automatically after being rebuilt on Board. Those are not toy demos.

Why This Changes the Picture for Designers Who Build

If you are building vibe-coded projects with AI, your current workflow probably involves one AI tool doing one thing when you ask it. That is already powerful. But the ceiling is real: a single agent loses context, forgets earlier decisions, and can only do one thing at a time.

Board points toward a different model. Imagine describing a feature you want, and instead of one agent grinding through it sequentially, several agents split the work. One handles the data layer, one writes the UI logic, one runs checks, and they all report back to a shared thread you can read in plain English. You stay in the director's chair while the agents handle more of the execution.

The practical shift for designers is less about writing code and more about shaping workflows. Board treats tasks as structured objects with statuses, ownership, and history. That is basically a design brief, broken into steps, with accountability baked in.

What to Actually Watch in Board's Feature Set

A few specific things are worth understanding:

Claims and leases. When an agent picks up a task, it "claims" it for a set window of time, a lease. If it finishes early or runs long, it can renew or release. This prevents two agents from accidentally working on the same thing at once. For a designer managing a complex project, that guard against duplication is genuinely useful.

Two-way sync with Slack and GitHub. Board mirrors messages, reactions, and edits between itself and connected tools. If your project already lives in GitHub, issues and pull requests can flow into Board automatically. You do not have to change your whole setup to try it.

Scheduled follow-ups. An agent can tell itself to check back in five minutes, or any interval you set, without you having to prompt it again. That makes agents feel more proactive rather than passive. You set a task in motion and the agent keeps an eye on it.

Access grants. You can give different agents, or different humans, different levels of read and write access per channel. A security channel can be locked down to specific agents only. This matters as soon as you are working on anything beyond a solo experiment.

Compatible with the tools you may already use. Board works with Claude Code, Codex, LangChain, Mastra, and others. It is not locked to one AI model or harness, which is a meaningful choice given how fast that landscape is moving.

The Honest Limits

Board is explicitly experimental. Vercel is inviting developers to test it and give feedback, which means the surface area will change and some things will break. It also skews technical right now. Setting up channels, writing watchers, and connecting sync integrations requires comfort with a command line and some JavaScript configuration. That is a gap for designers who are still finding their footing with code.

The bigger open question is orchestration. Board handles coordination between agents beautifully, but someone still has to design the workflow: what agents run, in what order, with what instructions. That is actually a design problem, not a coding problem, and it is one that tools like this will increasingly put in front of people like you.

The groundwork Vercel is laying here is worth watching closely. Multi-agent coordination is not a distant future concept. It is running in production today, and Board is a window into what building with AI looks like when the ambition scales up.

vercelai agentsdeveloper toolsworkflowvibe coding

Sources

About VibeLab

VibeLab is a community for vibe coders. Discover vibe coded projects from around the globe, and post and host your own.