Featured Story
How Vercel Uses Notion to Scale Launches Across the Company


Combining versatility, acceleration, and excellence in both its name and product, Vercel gives developers a frictionless path from code to cloud. This same philosophy powers their internal operations through an AI-enhanced Notion workspace, driving 35% faster shipping and helping teams reclaim up to nine hours weekly per employee. Their unified system has resulted in 89% of employees reporting increased confidence in the quality of their shipped products.
Every Vercelian now builds software
Tom Occhino, Vercel's Chief Product Officer, came back from the 2025 winter holiday and told the company he'd spent countless hours over the break building with AI. He's an engineer by trade (one of the creators of React) and he said he felt like he'd pushed more code during those two weeks than he had in the prior three years at Vercel combined.
Vercel had already been building agents. Across teams, Vercelians had been quietly automating work, improving output, and finding ways to scale more efficiently with AI from the bottom up, without a mandate. So when Opus 4.5 released, it crystallized something the Vercel team had already been feeling. It was time to formalize that into a company-wide motion. The principle was simple: Every Vercelian now builds software.

How Vercel’s launch process became the foundation for everything
For many of Vercel's teams, Notion became the place where agents were built, connected, and deployed. A lot of the foundational work at Vercel started with their ShipOps team, which is responsible for keeping the company’s fast-moving release culture running without friction. Brian Emerick, a Technical Program Manager on that team (and self-described “Chief Notion Officer”) has been at the center of building out what’s become a fleet of agents, all orbiting the same core system: The Launch Calendar.
If you’re shipping something at Vercel, there are three requirements: Write docs, write a changelog, put it in Launch Calendar. It is a single Notion database that every team in the company uses to track what is shipping and when, and it serves as the communication backbone for every release, keeping engineers, marketers, and support teams in sync.
Simple in theory. But over time, the database has grown to include around 50 properties covering everything that all those people need to do their jobs. The team’s first stab at making that manageable was a Notion form with conditional logic. It helped, but it was still a form.
The next step was Ship, a Custom Agent that works conversationally in Slack to create a Launch Calendar entry. Instead of working through a form, you give it what you have, such as a name, a date, or a link to whatever work is ready, and it figures out the rest by flagging what you told it directly, what it inferred, and what it is still unsure about. It then creates an entry and shares back a link in the Slack thread. The goal is faster intake, and more launches in the database that might never have made it in before.

Ship-DX connects every launch to every team
Ship solved the creation problem, but getting a launch into the database was only half the battle. They also needed to make sure the right people across the company knew about each launch and could take action in the tools they already use.
That’s what Ship-DX does. When a new launch is created or updated in the Launch Calendar, ShipDX automatically creates a corresponding Linear issue for Vercel’s developer experience (DX) teams, including docs, community, and others. It pulls details from Notion, fills out ticket descriptions, sets due dates, creates sub-issues, and assigns everything to the correct team based on each task. It also pings the launch leader in a shared Slack thread to coordinate with the relevant DX teams on those issues. When things change in Notion, Ship-DX comes back and updates the Linear tickets to match, plus a comment explaining what changed and why.
To do it, Brian’s team built a Notion Worker that makes the GraphQL call directly to Linear to create issues from templates, then hands off to the Linear MCP for subsequent updates and comments. It’s a small example of the larger trend: Notion as the connective layer between tools, with Workers filling in gaps to make everything happen.
The same logic extends to closing the loop on launches that have already shipped. Ship Closer wakes up each morning, checks Vercel’s public changelog against the Launch Calendar, and tries to reason whether a given launch actually happened on the day it was supposed to. If confident, it marks the entry as shipped. If it’s not sure, it waits (it has a built-in grace period so it doesn’t spam the same person every day) and follows up in a Slack thread where all launch partners already have visibility. It fits a broader principle at Vercel: Work in public, so everyone with a stake in a launch always has the latest information. Ship Closer is a small agent with a specific job, and it does that job without anyone having to remember to do it themselves.

How GTM Engineering made Notion the brain behind their agents
Drew Bredvick runs GTM Engineering at Vercel, an applied AI team focused on building tools that help sales and marketing close more deals. Unlike ShipOps, Drew's team builds fully custom agents in code, deployed on Vercel's own infrastructure.
But the further they got into building, the more Notion kept showing up as the thing that made everything else work. Every agent the team built had five components: A trigger, tokens, compute, a prompt, and observability. Of those five, the prompt was the one that mattered most: the thing you refine constantly and the thing that determines whether an agent does its job well or poorly. And for a long time, those prompts lived in codebases.
When Nick Bogaty, Vercel’s CRO, had a question about what was driving the decisions of one of their sales agents, there was no clean answer. The logic was buried in GitHub, inaccessible to anyone who wasn’t an engineer, and impossible to update without a pull request, a review, and a full deployment. A change to agent behavior shipped like a bug fix, slowly and with a lot of unnecessary hands involved.
The team's solution was to move all user-configurable prompts into Notion. Now a sales leader can open a doc, edit the prompt directly, and the agent picks it up on its next run without requiring engineering involvement. The iteration cycle dropped from roughly a business day to five minutes, and the people who actually understand the business logic, not just the code, are the ones keeping agents sharp.
Workers let us connect directly to other tools' APIs and automate what used to be manual handoffs. Notion becomes the connective layer, and Workers fill in whatever gaps exist between your tools.

Notion as the connective tissue behind Vercel’s agent stack
Notion isn't the only place Vercel's agents live. But for Vercel’s teams, in different ways and for different reasons, it's become the thing that makes them work. Notion is the canvas where some agents are built, and the layer of knowledge and logic that keeps others honest.
The pattern that Brian's team established with ShipDX is spreading with similar agents for pricing, legal, and other partner teams already on the way. At a company that prides itself on shipping fast, that kind of momentum doesn’t happen by accident. It starts with knowing what to build first.


