Oral board preparation has a well-known paradox. The single most effective way to prepare a resident for the Certifying Exam is also the hardest thing for a program to provide at scale. Mock orals are valued by faculty and residents alike, yet almost no program offers them in anything approaching the volume residents actually need.
The reason is not that faculty are unwilling. The reason is structural. Traditional mock orals are built on a resource — faculty time — that does not multiply. When demand grows, the constraint stays fixed. That is what it means to say mock orals don't scale, and it is worth understanding precisely, because the solution depends on getting the diagnosis right.
The Scarcity Problem
A traditional mock oral requires a faculty member to sit across from a resident for the length of a case, and often longer. A single realistic oral board case runs twenty to thirty minutes, and a useful session — one that approximates an actual exam with feedback afterward — can easily consume an hour of a surgeon's time once preparation and debrief are included.
That hour is not free. Faculty time in a surgical department is among the most expensive and most contended resources the institution has. Every hour spent running a mock oral is an hour not spent in the operating room, in clinic, on research, or on the administrative work that already crowds an attending's calendar. The arithmetic is unforgiving: a cohort of ten residents, each needing even a modest number of practice sessions before the exam, quickly translates into dozens of faculty hours that simply do not exist.
The result is rationing. Most programs are able to offer a resident one or two formal mock orals in the run-up to the Certifying Exam, if that. The rest of the preparation — the repetition that actually builds fluency — is left to informal practice among peers, which is valuable but ungraded and inconsistent. Scarcity, in other words, is not a bug in how programs run mock orals. It is a property of the format itself.
The Consistency Problem
Even when faculty hours can be found, a second problem appears: grading is hard to keep consistent. An oral board exam is a subjective, conversational assessment, and different examiners notice different things. One attending weights the differential diagnosis heavily; another cares most about the operative plan; a third is watching how the resident handles a deliberately vague presentation. All three are right, and none of them produces a score that maps cleanly onto the next.
This inconsistency matters more than it might seem. Residents use feedback from mock orals to decide what to fix next, and feedback that shifts depending on who happened to run the session is feedback that is hard to act on. A resident told one month that their workup was thorough and the next that it was underdeveloped may simply be receiving two examiners' standards rather than a signal about their own performance.
The Cost of Drift
Consistency is not a matter of faculty effort. It is a matter of structure. Standardized rubrics, training, and shared scoring anchors can narrow the variance, but they add another layer of faculty time to an already scarce process. The tension is real: the very steps that would make grading fairer also make mock orals more expensive to produce.
The Volume Problem
Fluency in the oral board is built on repetition. A resident who has talked through five cases speaks differently from one who has talked through fifty. The difference is not knowledge — both residents may know the material cold — but the practiced ability to structure a presentation, respond to interruption, and commit to a plan under time pressure. These are skills that only repetition produces, and the relevant unit of progress is simply the number of cases spoken out loud.
Traditional mock orals cannot supply that volume. Even a well-resourced program that runs a handful of formal sessions per resident is delivering a fraction of the repetition that oral board performance actually rewards. The gap between the volume a resident needs and the volume a program can provide is the core of why mock orals don't scale: the need is nearly unlimited, and the supply is sharply bounded.
What makes this gap consequential is that oral board performance is not evenly distributed. It is the difference between a comfortable pass and a narrow fail, and a narrow fail carries real costs — delayed board certification, a retake, and the professional and personal weight that comes with it. Programs feel this acutely, because they are accountable for their residents' outcomes.
What Scaling Actually Requires
If mock orals are going to scale, the scaling has to come from somewhere other than faculty hours. The options are limited. A program can schedule more sessions and accept the cost, which works only at the margins. It can shift some practice onto peers, which increases volume but sacrifices the calibrated, expert feedback that makes a mock oral valuable. Or it can change the mechanism of delivery — decoupling the act of practicing from the presence of a faculty examiner.
The last option is the one that actually changes the arithmetic. Once practice no longer requires a faculty member's live time, the constraint on volume disappears. The unit cost of one more case drops toward zero, and the question shifts from "how many sessions can we afford" to "how many sessions does each resident need." That is what scaling actually requires: not more hours, but a different mechanism.
The Question-Bank Precedent
This is a familiar pattern. Written board preparation went through the same transition a generation ago, when question banks made it possible to work through thousands of questions on demand, without an instructor present. The result was not less instruction — it was more practice, layered on top of the teaching that faculty already provided. The same logic applies to the oral board, once the delivery mechanism no longer depends on live time.
How AI Practice Fits In
An AI examiner changes the delivery mechanism without changing the exercise. A resident can still be presented with a clinical scenario, asked to talk through the workup and management, and pressed on their reasoning — the same structure as a live mock oral. What changes is that the examiner is always available, at any hour, for as many cases as the resident wants to run.
The consequences for a program are direct. Volume is no longer gated by the faculty calendar. A resident who needs to work through a weak area — trauma, hepatobiliary, vascular — can run ten cases on it in a week without anyone having to schedule anything. The feedback loop tightens, because every session produces a grade and a specific read on what to fix, which means the resident arrives at the next faculty-led session having already corrected the basics.
What Faculty Time Is Still For
None of this replaces faculty. It reframes what faculty time is for. Rather than spending an hour running a resident through a case they could have rehearsed on their own, a faculty member can spend that hour on what AI cannot do: calibrating a resident's judgment, sharing clinical nuance, and reviewing the handful of sessions that genuinely need a senior clinician's eye. AI practice multiplies the volume of repetition; faculty practice supplies the calibration and depth on top of it. The two are complements, not substitutes.
For a program director, this is the part that matters. The question is no longer whether you can afford to give residents enough practice — it is whether you can give them enough, consistently, without spending more faculty hours than you already do. That is the difference between mock orals that are rationed and practice that is simply available.
The Bottom Line
Traditional mock orals are high-value and inherently scarce. The format demands live faculty time, and live faculty time does not multiply, which is why mock orals don't scale on their own. Grading is inconsistent across examiners, volume is capped by the calendar, and residents repeatedly arrive at the exam with less oral rehearsal than the format rewards.
The fix is not to abandon mock orals — it is to change the delivery mechanism so that practice no longer depends on scarce hours. AI practice multiplies volume without multiplying faculty time, and it does so without replacing the faculty judgment that makes the feedback meaningful. The result is a cohort that gets both the repetition it needs and the calibration it could only get from its attendings.
Scaling oral preparation is not a matter of trying harder or scheduling more. It is a matter of recognizing that the constraint was never effort — it was the format. Remove the constraint, and the practice volume a resident needs becomes a decision about the resident, not about the calendar.