Based on reporting by The Information and CNBC.
OpenAI lost the top two leaders of its enterprise-sales organization in roughly eight days — precisely as enterprise revenue overtakes consumer and an IPO looms on the horizon.
What Happened
Reporting from The Information confirmed on August 21 that Kaylin Voss, OpenAI’s vice president of sales for the Americas, has resigned — roughly a week after her boss, chief revenue officer Denise Dresser, left the company after less than a year in the role. Voss reported directly to Dresser, which means the top two leaders of OpenAI’s enterprise-sales organization exited within approximately eight days of each other. Both came to OpenAI from Salesforce; Dresser joined in December 2025 and Voss is now returning to Salesforce, where she is publicly welcomed back by Marc Benioff.
One crucial fact keeps this from being an empty-chair story: on the same day it announced Dresser’s departure, OpenAI named a new chief revenue officer — Dali Rajic, the former president and chief operating officer of the security company Wiz. There is a named successor at the top of the revenue organization. What is unsettled is the layer directly beneath him, with the VP running Americas sales now gone too, and Rajic himself still new to the organization and its enterprise motion.
Context from the same reporting: roughly a dozen senior leaders have exited OpenAI in 2026, a group that includes names such as Brad Lightcap and Fidji Simo among the year’s departures. Others on the sales team are reportedly weighing exits. Those figures come from people familiar with the situation rather than from OpenAI itself, and a raw count of departures is not a measure of business health — but the pattern is the backdrop against which this particular pair of exits lands.
The key insight: The risk here is not a leaderless vacuum — OpenAI installed Dali Rajic as CRO the same day Dresser left. The risk is handover and continuity: a new revenue chief, still onboarding, taking over an enterprise engine at full sprint, with his second-in-command now also gone. Continuity is precisely what a fast-scaling sales motion is most fragile about.
Before the Analysis: What This Is Not
Executive churn at a hyper-scaling startup is common. A departure or two is not, on its own, evidence of trouble, and the most important single fact about this transition is one that is easy to skip past: Dresser tripled the size of the sales team during her tenure. The organization she leaves behind is larger and more built-out than the one she inherited — not hollowed to the studs. The motion she built outlasts the individuals who built it.
The alarming-sounding numbers are soft. “Roughly a dozen departures” and “others considering leaving” are sourced to people familiar with the situation, not disclosed figures. Most importantly, there is no evidence the enterprise business is deteriorating — by the same reporting it is reportedly still accelerating. Reading a leadership reshuffle as the business breaking gets the causation backwards. And Voss returning to Salesforce, where she has roots, is as much a personal and career decision as it is a signal about OpenAI. The instability-at-the-IPO-doorstep framing that follows is our analysis and that of outside observers — a soft signal, not a hard one.
The Structural Read
Map of AI — GTM Layer
The GTM Engine Wobbles at the Worst Possible Moment
OpenAI’s most valuable revenue line is now the one whose leadership just walked. Enterprise recently overtook consumer by revenue and is accelerating. The people who built that acceleration are, for now, harder to find on the org chart than the growth itself.
Execution risk on the high-value revenue line. OpenAI’s enterprise business recently overtook its consumer business by revenue, a milestone driven in part by the GPT-5.6 launch and the stickier, higher-margin contracts that come with it. Enterprise sales — with long procurement cycles, executive relationships, and complex deployments — is the revenue line a maturing AI company is supposed to build its future on. Losing the two leaders who built that motion is a risk concentrated precisely on the line the company’s story now depends on. Rajic is a named, capable successor, but he is onboarding into an organization at full sprint, without the VP who ran its largest geography. That is not an empty-chair gap; it is a continuity and transition problem. The distinction matters, but so does the problem.
OpenAI’s enterprise revenue growth and the crossover moment are covered in more depth in our analysis of how enterprise revenue exceeded consumer ChatGPT.
The Salesforce round-trip. Both Dresser and Voss came from Salesforce — the archetype of enterprise-software go-to-market — and Voss is returning there directly. That loop is a reminder of something structural: the talent who know how to sell complex software to large organizations are scarce, highly mobile, and hard for a model lab to retain. OpenAI is still, in a meaningful sense, learning to be an enterprise vendor rather than a consumer-chatbot company that also sells to businesses. The people who knew how to do that are now back at a company that invented the playbook. This is the deeper pattern beneath the personnel news, and it connects directly to the competitive dynamics we mapped in the OpenAI–Anthropic enterprise consulting war.
Churn at the IPO doorstep. Roughly a dozen senior departures across 2026, set against preparations for a potential IPO, is a stability and governance optics problem regardless of the underlying causes. Public-market investors do not just price growth rates — they price the durability and depth of the management team executing on that growth. A new CRO who is still onboarding, a departed VP beneath him, and a broader pattern of 2026 exits is the kind of profile that shows up in an S-1 and gets underwriters asking hard questions. The analysis of what this means for OpenAI’s path to the public markets is in our OpenAI IPO 2027 piece.
Three Implications
IMPLICATION 1 — Rajic’s Onboarding Is Now the Critical Path
A new CRO inheriting an enterprise engine at full sprint, without the VP who ran its largest geography, faces an onboarding problem that is also a business risk. The speed at which Rajic builds relationships, learns the pipeline, and fills the Americas VP seat will determine whether the growth Dresser built continues to compound — or plateaus while the organization reorients. This is the number to watch in OpenAI’s next revenue disclosures, whenever they come.
IMPLICATION 2 — Enterprise GTM Talent Is a Structural Moat OpenAI Does Not Yet Own
The Salesforce round-trip is not a coincidence. Enterprise software go-to-market is a learned discipline with its own culture, playbooks, and networks — and Salesforce is its academy. Until OpenAI develops enough institutional GTM knowledge to retain the people who build it, it will remain dependent on importing that knowledge from legacy software companies, with the attrition risk that entails. Anthropic and others are competing for the same scarce pool.
IMPLICATION 3 — The IPO Narrative Now Has a Personnel Chapter
Any prospectus OpenAI files will have to account for 2026’s senior departure count. Underwriters and institutional investors price management stability alongside revenue growth, and a pattern of roughly a dozen exits in a single year — including the two leaders of the company’s fastest-growing revenue line — is a disclosure and optics challenge, independent of whether the business itself is healthy. That challenge is manageable; it is not invisible.
The Bottom Line
OpenAI’s enterprise business is reportedly still growing, Dresser tripled the team before she left, and Dali Rajic is a named successor — not a vacancy. None of that is in dispute. What is also true is that the two people who built OpenAI’s enterprise-sales engine exited in the same eight-day window, to the company they came from, at the precise moment enterprise revenue is the most important line on the model and a potential IPO makes management continuity a public-markets question. The enterprise GTM motion Dresser built is larger than she was — the motion outlasts the individuals. Whether it is also more durable than this particular leadership transition is what the next two quarters will answer.
Sources: The Information — Kaylin Voss resignation · FourWeekMBA — OpenAI Enterprise Revenue Exceeds Consumer · 91,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.









