The No. 2 Exit: Why Boards Underestimate Deputy Departures

In the space of one week this July, two of the most consequential seconds-in-command in American technology announced their exits. Neither departure will show up in a CEO turnover index. Both may matter more than most CEO changes will this year.

On July 9, Fidji Simo — OpenAI's No. 2 executive and the leader of its consumer business — told staff she was stepping back from her full-time role and transitioning to part-time advisory work, after a medical leave that proved longer and harder than expected. Days earlier, Sumit Rana, president of Epic Systems, the electronic health records company whose software anchors most of America's largest health systems, announced he would step down on August 14 after more than two decades, citing his father's death and family obligations in India.

Two different companies, two different reasons, one identical structural consequence: a succession picture that looked settled on Monday looked open by Friday.

The Load-Bearing Seat

Boards, investors and the business press track CEO transitions obsessively. Weekly turnover indexes count them; share prices move on them. The No. 2 seat gets no such scrutiny — and that inattention is a measurement error, because in many organizations the deputy is the load-bearing executive.

Consider what Simo actually held. When she joined OpenAI in May 2025 from Instacart, where she had been CEO through its IPO, the company restructured its reporting lines around her: the chief operating officer, chief financial officer and chief product officer all began reporting to Simo, freeing Sam Altman to concentrate on research, compute and safety. She was not an understudy; she was the operating system. TechCrunch reported that she was widely viewed as a candidate for even greater responsibility if and when OpenAI went public. Her departure, in that framing, is not the loss of a lieutenant — it is the removal of a keystone at precisely the moment the structure was being built to bear IPO-scale weight.

Rana's case is the succession question in even purer form. He joined Epic in 1998 as a software developer and rose to president in 2024. Epic's founder and CEO, Judy Faulkner, is 82. Rana was, by most industry accounts, the most visible internal candidate to eventually succeed her. Epic has said it is not naming a new president. The dominant company in American health IT — privately held, famously founder-controlled — now has a succession picture with no publicly legible answer. Rana himself, in his farewell note to employees, called Epic "the best thing that ever happened to me." The warmth of the exit does not change its arithmetic.

Why Boards Under-Index the Deputy

There are structural reasons the No. 2 departure gets less attention, and none of them survive scrutiny.

The disclosure gap. CEO changes at public companies trigger filings, press releases and analyst calls. Deputy departures often surface through internal memos and secondhand reporting. Less disclosure reads as less importance — but disclosure requirements track legal formality, not operational reality.

The figurehead fallacy. External stakeholders interact with the CEO, so they overweight the CEO's importance to daily operations. Inside most companies, the deputy is the person who converts strategy into roadmaps, holds the customer and board relationships that don't make headlines, and manages the executives who manage everything else.

The bench illusion. Boards tend to treat a named deputy as evidence that succession planning is complete. It is closer to the opposite: a single visible successor is a single point of failure. When Rana leaves, Epic's succession plan does not shrink by one option — it reverts to an open question. When Simo leaves, Altman is not merely down one executive; he is re-litigating an organizational design that was built around a specific person.

The Second-Order Effects

Deputy departures also cascade in ways CEO exits often don't. A CEO transition typically comes with a plan — an announced successor, a transition timeline, an orderly handoff. A deputy exit tends to come without one, which means the vacated portfolio gets redistributed among executives who already have full-time jobs. The redistribution itself becomes an informal succession contest, with all the political energy that implies.

There is also a talent-market signal. High performers inside an organization watch who ascends and who exits. When the presumed successor leaves, the internal message — intended or not — is that the path upward is either blocked or unclear, and the strongest people start taking calls from recruiters. The deputy vacuum can become a talent drain before the org chart is redrawn.

What Well-Run Boards Do Differently

The practical prescriptions are unglamorous but specific. Treat the deputy seat with the same succession discipline as the CEO seat: a named emergency plan, at least two developed candidates rather than one, and a standing board conversation about what happens if the No. 2 leaves — not just the No. 1. Stress-test organizational designs that concentrate reporting lines in a single executive, as OpenAI's did; concentration buys coherence in good times and fragility in transitions. And in founder-led companies with long-tenured leaders, resist the comfort of a single heir apparent, because the actuarial math on an 82-year-old founder and a sole successor was never as safe as it looked.

Neither Simo nor Rana left over strategy or performance — health and family, the two reasons no retention package can address. That is precisely the point. The unplannable departure is the one succession planning exists for, and it applies to the deputy seat every bit as much as the corner office.