Grit · 5 Oct 2026 · From the week of 5 October
How Vercel Is Rethinking Go-to-Market With AI | COO Jeanne Grosser
Listen on Apple Podcasts Show on Spotify Episode page
In brief
Kleiner Perkins partner Joubin Mirzadegan interviews Jeanne DeWitt Grosser, COO of Vercel, who owns all of go-to-market (marketing, sales, field engineering, support and partnerships) after about a decade at Google and over nine years at Stripe, where she left as chief business officer. They cover how she separates product-market-fit problems from go-to-market problems, the funnel 'operating model' she builds on arrival to test the founder's revenue target, how she timed a major sales reorganisation, and how to work with a technical founder. Her central argument is that AI is changing the shape of go-to-market teams. At Vercel, agents resolve 91% of support volume and handle all inbound qualification, and quotas have been raised every six months. She expects leaders to be judged on headcount-to-revenue ratio rather than the size of their org. Joubin adds his own views on why CROs get fired, inherited founder forecasts, and headcount as a vanity metric.
For founders
- Before blaming sales for slow growth, put your product manager and lead engineer through about ten back-to-back customer conversations; Vercel did this and concluded it did not yet have full enterprise product-market fit.
- Have a new go-to-market leader build a full-funnel operating model by segment and geo against your revenue target, so missing pipeline shows up months before the number is missed.
- Grosser treats the founder relationship as a 'professional marriage': she insisted on a scheduled weekly one-on-one with a founder who did not hold them, and found long dinners to be a major unlock.
- Grosser expects AI to let functions such as support, BDR and RevOps stay much leaner, and planned to have her leadership team work through the 'theoretical end state' headcount of each function at an upcoming offsite.
- When a founder repeatedly flags small details, Grosser treats a recurring pattern as a sign of a systemic problem and argues openly about priority when the founder sees a P0 and she sees a P2.
For revenue leaders
- Vercel resolves 91% of support volume and all inbound qualification with in-house agents, and much of its outbound is done by an agent; the remaining support and BDR staff work on harder problems such as fixing product issues and complex enterprise prospecting.
- Vercel has raised sales quotas every six months and its headcount-to-revenue ratio keeps rising, so it can hit bigger targets without adding a rep per increment of revenue. Grosser illustrated this by saying that instead of $1M per head (10 more reps for $10M), you can raise per-head capacity, say to $1.5M. She credits product maturity and market readiness for part of this, not only AI.
- Use AI to extract objections from call transcripts, score how well reps handled them, and automatically file product requests when an objection reflects a product gap.
- Change sales structure and compensation only at the start or middle of the year, and plan for a near-term dip after a big reorganisation.
- Evolve go-to-market every six months toward more specialisation (technical sales roles, verticals such as e-commerce) instead of doing big reorgs.
What was said 32, most useful first
Vercel has raised sales quotas every six months, lifting capacity per rep instead of adding heads for each new increment of revenue.
Grosser said quotas have gone up every six months of her roughly year and a half at Vercel, and its headcount-to-revenue ratio keeps rising meaningfully. She credited part of this to product maturity and market readiness, not only AI efficiency gains. To illustrate, she said the traditional approach is about $1M per head, so $10M more means 10 more reps, whereas now you can, for example, take the $1M per head to $1.5M and not need 10. She said reps accept this because exponential growth makes most people 'come along for the ride', though not everyone loves her at the start of each half.
“I want 10 million, but I can take the million bucks to a million five. And so now I, you know, don't need 10.”
Listen on Apple Podcasts Sales team, hiring & comp Link to this
To settle whether slow growth is a product problem or a go-to-market problem, Vercel ran its PM and lead engineer through 10 back-to-back customer conversations, and concluded it lacked full enterprise product-market fit.
Grosser said no one has invented a test for product-market fit, so the question of product versus go-to-market keeps coming up. At Vercel she picked 10 large existing enterprise customers that used Vercel but not yet in front of their main site, and had the product manager and lead engineer hear all 10 in a row to build pattern recognition. Afterwards everyone agreed there were product gaps to close before the product would be much more easily sellable. She noted that some enterprises still buy before full fit, because a visionary buyer will always work through the pain.
“So we're like, great, let's go get 10 of those. Let's get the product manager, lead engineer, and let's run them through conversations with all 10 in a row, so they can get pattern recognition.”
Grosser's funnel model showed Vercel's board, a few weeks after she joined, that pipeline gaps would make the target unreachable by Q3, so she made a CMO her first leadership hire.
She said the business would look fine for two quarters, but not enough pipeline had been created in Q4 before she arrived, or in Q1, to support the planned ramp. Showing the full-year trajectory early prevented two good quarters from masking a later 'cliff'. Because top-of-funnel was clearly the top issue, she hired the CMO first, within 90 days. By Q4 enough levers had changed that Vercel 'just barely' beat the number, which was kept rather than lowered.
“The first one I hired was the CMO, because it was immediately obvious. Our top issue is marketing, pipeline, top of funnel. Got him in within 90 days”
Listen on Apple Podcasts Pipeline & demand generation Link to this
Change a sales organisation only at the start or middle of the year, and get the executive team to accept a near-term dip in advance.
Grosser said reps' livelihoods depend on quota attainment, so you can't change everything on a random Tuesday. At Vercel she made major go-to-market changes four and a half months in, faster than she naturally would have, because waiting until six months would have landed mid-Q3. The executive team agreed beforehand to accept a possible near-term trough for the upside. About six months after the changes she was again not sleeping well, because not everything could be operationalised cleanly; roughly six weeks after that, the change looked right.
“it's easiest to change a sales team sort of at the halves.”
Listen on Apple Podcasts Sales team, hiring & comp Link to this
A new CRO who adopts the founder's forecast without knowing their pipeline data will be fired within a quarter.
He described an early-stage CRO, two to three months in, committing to $10M for the year. The CRO couldn't state current pipeline and was counting on an unreleased self-serve product and a $1M big-tech deal that had only had a first meeting. Joubin's view is that founders are default optimists who 'do not know how to forecast a business' and need a big number to justify their valuation. His advice was to re-forecast from data, not forecast a PLG product that doesn't exist yet, and push the product team to fill the gap, because the product leader won't be fired first.
“I would not forecast the PLG product that doesn't exist yet.”
Joubin Mirzadegan relayed advice from Rubrik's CEO to deliberately flood a company with demand early to learn how fast it can grow.
Joubin, whose own company planned to grow headcount 6x in nine months, said Rubrik's CEO pushed him to grow even faster. According to Joubin, the CEO's first four sales hires at Rubrik were BDRs, meant to 'suffocate the company in demand'. When they later hit an air pocket it didn't matter, because the goal was to become relevant as quickly as possible. Joubin framed this as an interesting counterargument to running lean.
“the first four hires that I made were in sales were four BDRs. And I just like literally tried to suffocate the company in demand, in pipeline.”
Listen on Apple Podcasts Pipeline & demand generation Link to this
Joubin Mirzadegan warns that headcount growth is a seductive proxy for success, and instead challenges new leaders to run the highest-performing team with the fewest people.
He said bragging about multiplying headcount, like valuations, can become detached from the actual business. It can also mask weak product-market fit by willing growth into existence through hiring. When a new leader asked to be pushed harder, he set that goal of the highest-performing team with the fewest people.
“can you be the highest performing team with the least amount of people?”
Listen on Apple Podcasts Hiring & team building Link to this
Grosser evolves the go-to-market structure every six months toward more specialisation instead of doing large reorgs, and Vercel recently carved out an e-commerce vertical.
When she joined Stripe there were only account executives and account managers, with no technical sales roles or BDRs. As Stripe moved upmarket into migration conversations and regulated industries, it added specialisation about every six months. Her stated aim was 'we'll never reorg', just take the natural next step each time. At Vercel, selling to e-commerce differs enough from selling to tech companies that it had just created an e-com vertical. Her signal for the next step is when a generic sale stops working in a newly opened part of the market.
“If I'm doing my job right, we'll never reorg.”
Listen on Apple Podcasts Sales team, hiring & comp Link to this
Cognitive brain-teaser tests for sales candidates may have zero or even negative correlation with job performance.
He said some technical founders put sales candidates through puzzle-style tests. He compared this to quizzing an engineer on MEDDPICC or quota capacity. Grosser replied that salespeople joining engineering-led companies need to know what they're signing up for.
“this has maybe zero, if not negative correlation to their performance on the job.”
Listen on Apple Podcasts Sales team, hiring & comp Link to this
Vercel resolves 91% of its support volume with an AI agent, and the remaining human support team finds and fixes real product problems instead of working through tickets.
Grosser said Vercel gets a lot of support volume and 91% of it is resolved by an agent. The remaining support team, which is technical, now focuses on legitimate product problems, works closely with engineering, and in some cases writes pull requests to fix issues itself.
“Vercel gets a lot of support volume, as you'd expect, and 91% of it, we resolve with an agent.”
At Vercel all inbound qualification and much outbound is done by agents, so the smaller BDR team does complex enterprise prospecting and tests new hypotheses.
Grosser said the BDR function is 'way, way leaner' than at Stripe because agents handle inbound qualification and much outbound. The remaining BDRs either do more complex prospecting into enterprise or are used for hypothesis testing, for example on a new persona or new messaging.
“All the inbound qualification is done with an agent.”
Listen on Apple Podcasts Pipeline & demand generation Link to this
Vercel's RevOps team is leaner because a data science agent answers questions that would previously have gone to RevOps.
Grosser said Vercel has a data science agent you can ask anything you would have sent to RevOps. As a result, she described the RevOps team as leaner and more efficient than teams she had in prior roles.
“We have a great data science agent that, like, I mean, you can ask it anything you would have sent to RevOps.”
Grosser has product managers, engineers and the founder watch her deliver the pitch, then asks them whether she described the product correctly.
She said she asks Vercel's founder after every sales call they take together whether she got anything wrong in the pitch, and tries to absorb how he talks about the product. Her view is that if the founder and product managers agree the product is being described properly, that also tells you something about whether the problem lies in go-to-market or in the product.
“get your product managers or engineers in front of customers and watch me deliver the pitch.”
Vercel uses AI to extract objections from sales calls, score how well they were handled, and automatically file product requests when an objection reflects a product gap.
Grosser described analysing call transcripts for patterns as a way to get product and go-to-market to agree on what is and isn't working. Objections that can't be handled because the product lacks something are automatically submitted as product asks in Vercel's internal GTM feedback tool.
“we proactively extract objections from calls and score the degree to which we've handled them.”
Grosser's first move at both Stripe and Vercel was an 'operating model': full-funnel math by segment and geo showing what has to be true to hit the founder's revenue number.
At Vercel this is called the lead-to-revenue model. She said the spreadsheet lets everyone judge honestly whether the target is achievable, based on history or on assumptions they are willing to make, and later shows which driver caused a hit or a miss. At Stripe she owned new business, which was the gap between a 'nice large round' target and the forecastable existing base. The model showed quickly that the target was unlikely, which led to a strategic discussion of options: invest in marketing (Stripe then had one marketer), invest in outbound, or move upmarket for a larger ACV. She said the data was messy at first, partly because she had changed the segmentation, so reasonable assumptions filled the gaps.
“it's your full funnel math by segment, by geo.”
CRO tenures are short largely because technical founders find it easier to blame sales than the product.
He said the average CRO tenure is under two years. In his opinion a big reason is that most Silicon Valley founders are technical and see the product as a reflection of themselves, so they 'shoot the sales leader', and marketing leaders too. Grosser said she totally agreed and had experienced it earlier in her career.
“it's easier to blame sales than it is to blame the product.”
Grosser picks a company to stay at for a decade using three criteria: values actually used to make decisions, a large market with natural adjacencies, and world-class engineers.
After a Series B startup that wasn't a good fit, she worked out why she had stayed at Google for 10 years and used those reasons to choose Stripe, where she stayed over nine years. She applied the same criteria to Vercel. Its values are used in onboarding, all-hands meetings and performance reviews. Its market is horizontal and expanding into CDN, security and AI, and she believes 'every pixel on the Internet could theoretically be built on Vercel'. It is also a company of inventors. She said she found only about five companies that met the bar when she chose Stripe.
“One was it had very well articulated values that it truly used as like decision-making criteria.”
Marketing, sales and support should report to one leader because the human customer experience is a product in itself.
She said it is common for these functions to roll up to three separate leaders, which produces overlapping strategies rather than one integrated experience. She has repeatedly seen support and sales use different segmentations, which she considers unworkable, and has seen marketing spend misaligned with where sales puts people. Owning all of it makes aligning them her problem.
“one of the things companies don't nail is the human-based customer experience is a product just like the actual product.”
Owning both marketing and sales removes the option of blaming the other function for missed pipeline, which Grosser describes as another layer of accountability.
Grosser said she wanted to own enough of go-to-market that the revenue outcome truly rested on her. As COO over both marketing and sales, she can't say as CRO that pipeline shortfalls are marketing's fault, or the reverse.
“if we don't have enough pipeline, that's my fault.”
Grosser treats the founder relationship as a 'professional marriage': she insisted on a scheduled weekly one-on-one and found that dinners with the founder were a big unlock.
She agreed with Joubin that the relationship with the founder matters more than almost anything else in choosing a role. Vercel's founder did not start with a scheduled one-on-one with her, but she insisted on one because she takes nine-plus hours of customer meetings a day and can't catch him in the hallway. Two-hour dinners, which she says she should have started much sooner, feel less transactional and make room for important but non-urgent topics. She cautioned that what works may be very situational to each founder.
“We started, I should have done this way sooner, but getting dinner. That was like a big unlock too.”
Grosser suggests founders who dislike one-on-ones might be more effective if they kept Slack for real-time items and saved complex issues for in-person conversations.
Joubin said he is anti-one-on-one and prefers to be contacted ad hoc. Grosser said that dumping complex issues in Slack late at night forces executives to explain the context in a medium where it doesn't belong half the time. She put her suggestion tentatively: he 'might be more effective' with that split.
“I would purport that you might be more effective if you picked what belongs in slack in real time and like what might be like a more IRL substantive conversation.”
When a founder flags small details, Grosser looks for repeated patterns that point to a systemic problem, and discusses priority openly when they disagree.
Some one-off nitpicks can be acknowledged and deferred while she works on bigger priorities. If the same issue comes up repeatedly, for example bad grammar in emails three times in two weeks, she treats it as systemic. When the founder sees a P0 and she sees a P2, they either have a direct conversation and agree to disagree, or she checks whether it is really important, recognising that the founder is the founder.
“if you point out bad grammar in emails, three times in two weeks and I'm like, systemic problem”
Stripe rewrote its company values about every 18 months, and dropped 'efficiency as leverage' once leanness became a constraint on growth.
Grosser said Stripe's founders regularly reviewed which values to keep, which to discard, and what new behaviours needed a new value, because what got you here won't get you there. 'Efficiency as leverage' meant running very lean, finding ways to do more with less, and not throwing people at problems. After the first couple of years they realised the leanness was potentially rate-limiting, so that value was replaced.
“one of the things that Stripe was excellent at was rewriting its values every call it 18 ish months.”
Grosser leans toward growing headcount leanly because overhiring adds coordination cost, though she thinks Stripe probably could have grown revenue faster with more people at times.
She said that even when revenue isn't a problem, overhiring blurs job definitions, makes people step on each other's toes, and leaves more new people who don't know what they're doing than experienced ones. She finds lean growth more intellectually interesting because it forces hard prioritisation, automation and use of data. She admitted she has to consciously push herself on this, and that there were points at Stripe where more people probably would have meant faster revenue growth.
“You actually just slow the entire machine down with coordination cost.”
Listen on Apple Podcasts Hiring & team building Link to this
Grosser planned to have Vercel's go-to-market leaders define the 'theoretical end state' headcount of each function at an upcoming offsite.
She has 10-plus functions in go-to-market. She intends to classify each one: never needs more people than today, needs to add only a few a year, or scales non-linearly with revenue. Then the team would ask whether anything could change the shape of the curve for the functions that must scale. This was a planned exercise at the time of the interview, not one already completed.
“what is the theoretical end state of your function?”
Listen on Apple Podcasts Hiring & team building Link to this
Leaders will increasingly be judged by output and headcount-to-revenue ratio rather than by the size of the org they ran.
She described a candidate who currently runs a team and would join Vercel as an individual contributor. If she does the job right, with AI applied to it, Vercel would never hire a team under her. Grosser said many people will have to stop valuing themselves by 'empire building' and suggested people 'maybe' will start asking about headcount-to-revenue ratio, which she thinks might be healthier.
“people are gonna start asking, what was your headcount to revenue ratio rather than did you manage a 1,000-person org?”
Listen on Apple Podcasts Hiring & team building Link to this
Vercel requires everyone entering sales to start as a BDR, often hires former technical founders, and moved as many BDRs into solution architect roles as into AE roles.
Grosser said Vercel runs internal mobility twice a year with a formal process. Many of its BDRs were technical founders whose companies failed and who found they loved customer discovery. They were willing to start as BDRs because that is the required entry point into sales at Vercel. Vercel also hires former founders into other go-to-market roles.
“our BDR function, we graduated as many to be account executives as to be solution architects”
Listen on Apple Podcasts Sales team, hiring & comp Link to this
Joubin Mirzadegan objects to salespeople sending AI-generated summaries, arguing that synthesising what actually matters is the selling job itself.
He described reps running meeting notes from Granola through Claude and sending the output to the team. He also said every AI-made deck now looks the same. He told his team that if he doesn't want to read these, customers definitely don't. Grosser added her own complaint: AI-written call prep docs instruct her in an unnatural tone, such as telling her not to sell on an exec-to-exec call.
“your job is to synthesize what actually matters.”
AI has made Vercel's hypergrowth feel digestible, unlike when Stripe doubled headcount in a year during COVID.
Vercel grew from about 600 to just over 800 people in her roughly year and a half there, while revenue grew 'well north of triple digits'. She said she picks companies where revenue grows roughly exponentially while headcount grows roughly linearly, as at Google and Stripe. She attributes part of the difference to AI giving many roles multiple-times productivity.
“there are a bunch of things where like, you know, the equivalent human is getting NX productivity.”
Vercel expects every new employee to ship something in their first week, as part of a culture of building in public and iterating.
Grosser said Vercel's mantra is 'you can just ship things' and that the ability to ship can come from anywhere in the org. Because the company's roots are in open source, work happens in public, people comment on each other's work, and you have to be comfortable with feedback. Another value, 'iterate to greatness', favours acting and iterating over contemplating for a month. She said the result is autonomy paired with accountability.
“when somebody new joins Vercel, you're expected to ship something in your first week.”
An operator's job is to build systems that produce predictable output, so she tries to keep her own feedback to her team thematic and predictable.
She said she tries to make her pings about things her team already knows she cares about, rather than a constant flood of whatever is on her mind. For her the predictable output is revenue, so she actively weighs whether something that bugs her is a P0 for the next evolution of the system she is driving.
“An operator's job is to build systems that produce predictable output, basically.”
Grosser took three days off between Stripe and Vercel, and during her hardest early changes she wished she had taken more time to recharge.
She left Stripe on a Thursday and started at Vercel the following Monday, after over nine years of hypergrowth. Making substantial, hard changes as a newcomer drained her, and she said at points she should have taken time off. She added that she slept better than she had in a year on the day she announced the changes.
“there were definitely points through that window where I was like, man, I should have taken time off because my battery did not get recharged.”