Inside Vercel's agent-run go-to-market, from the COO's chair
Key points
Key takeaways from the Grit interview by Kleiner Perkins with Vercel COO Jeanne DeWitt Grosser, published October 2026:
Support on agents. Vercel resolves 91% of its support volume with an AI agent; outbound and all inbound qualification are likewise agent-run. The remaining support team now hunts genuine product problems, working closely with engineering — or writing the pull requests themselves — instead of churning through tickets.
The COO's remit. Grosser owns the entirety of go-to-market: marketing, sales, field engineering, support and partnerships. "If it touches a customer or makes a dollar, it's my job," she says — including a data science agent that answers anything staff would previously have sent to RevOps, and sales quotas that have been raised every six months.
Testing enterprise product-market fit. Unsure whether Vercel's slower enterprise growth was a product or a go-to-market problem, Grosser took ten large existing enterprise customers, sat their product manager and lead engineer through conversations with all ten in a row, and built pattern recognition together. The conclusion: full enterprise product-market fit was not yet there, and a defined set of product improvements was needed to make the platform easily sellable.
Objections as product signal. Vercel proactively extracts objections from sales calls with AI and scores how well they were handled. When an objection is unhandleable because of a product gap, the system automatically files a product ask in the go-to-market feedback tool.
The operating model. On joining a company, Grosser builds an "operating model" — the maths behind what must be true to hit the founder's number. At Vercel everyone knows the "lead to revenue" model: full-funnel maths by segment and geography, so the company can see which underlying driver explains any miss. At Stripe the same exercise showed a chosen top-line number was not attainable, enabling a strategic conversation about marketing spend, outbound and moving upmarket.
Beating the founder's forecast. Founders are default optimists who cannot forecast a business, Grosser warns; a CRO inheriting a founder's forecast without pipeline data will "get fired in a quarter". Her advice: refuse to forecast a product-led growth offering that has not shipped, work backwards from real data, and put the pressure back on product to fill the gap.
Why leave Stripe. Grosser spent over nine years at Stripe — from roughly 400 people to 8,000, and revenue from about $100M to billions — leaving as chief business officer, with three days off before joining Vercel, where she had already advised for 18 months. "You join a company like Vercel because you believe you're going to be part of building something more generational," she says: her goal is to write the go-to-market playbook for this era.
Read more: Grit — Kleiner Perkins (YouTube)
