The New Math of Admissions
A perfect storm is brewing in the high-stakes art of admissions: between surging applications, growing applicant use of generative AI, inflated GPAs, and overworked admissions teams, the system that we use to choose students is flooding across both undergraduate and graduate programs.
It's 11:47 p.m. on a Tuesday night. A PA program director closes her laptop after finishing another Zoom interview. She opens her browser to a familiar mess: spreadsheets, personal statements, and a dozen tabs for GPA lookups and prerequisite checks. Her child sleeps in the next room. The admissions chair has already pinged the committee twice this week: "We're behind."
This is the new reality. Applications keep rising. Available seats don't. The math simply doesn't work anymore.
The Impossible Task
What used to be a slim folder is now a digital maze. Today's application includes transcripts from multiple schools, recommendation letters, logged hours, test-optional contexts, video interviews, and portfolios. And let's be honest—more and more writing that's been polished by AI.
Choosing a great class used to be a joyful challenge. Now it's become an impossible task. We're asking humans to maintain perfect attention, fairness, and memory across months of life-changing decisions. It's inhuman work.
The manual process doesn't just struggle—it creates its own problems. Ask any committee member about how different reviewers score candidates, and you'll get a knowing smile. One loves "grit," another wants polish. Morning interviews feel different than afternoon ones. Two candidates thirty minutes apart get two completely different versions of us: one energized, one exhausted.
Research confirms what we already know: fatigue and context quietly change outcomes. In admissions, these small drifts become unofficial policy without anyone voting for them.
The Opacity Problem
Closed-door meetings make things worse. When a dean asks, "Walk me through why we chose Candidate B over Candidate A," we scramble for notes and half-remembered conversations. The path from evidence to decision gets lost in scattered PDFs, email threads, and interview notes buried in personal drives.
Even when everyone acts in good faith, the process can't explain itself. Not to the institution, not to the public—not at the level of rigor these decisions deserve.
Meanwhile, faculty and staff burn out. Evenings bleed into weekends. The workload moves into personal time because where else can it go? We rarely talk about the ethical problem here: the more we rely on after-hours heroics, the more we normalize a system that only works when people quietly donate their time.
That hidden subsidy makes for a fragile foundation.
Building Better
If we designed an admissions process from scratch today, what would we demand? Here's one vision, built on five pillars.
1 - Consistency
First, consistency that reduces human "noise" without eliminating human judgment.
2 - Transparency
Second, transparency—the ability to show our work and reconstruct how we weighed evidence, with records we can defend.
3 -Visibility
Third, visibility across interviews and committees so closed doors don't become closed systems.
4 - Usability
Fourth, usability—tools that reduce friction for the admissions team, instead of adding clicks, load times, and bureaucracy.
5 - Feedback Loops
Finally, feedback loops—processes that don't end at enrollment but track the quality of decisions based on post-matriculation performance, including persistence, remediation needs, board pass rates, and graduation.
Some people argue that formalizing the process risks making it sterile. But consistency doesn't make committees robotic—it makes them trustworthy. Structure actually protects the meaning in subjective judgments. When guidelines are clear and evidence is organized, the human insights that truly matter become easier to see and defend.
Why Now
This moment is urgent not just because of volume, but because our traditional inputs are changing with AI assistance. Essays, grades, and even interview responses now exist in an environment saturated with AI assistance. We can complain about it, or we can redesign for it.
This moment is urgent not just because of volume, but because our traditional inputs are changing with AI assistance.
The path forward isn't cynicism or nostalgia. It's a practical rebuild around signals that can withstand assistance and connect to the outcomes we actually care about.
A committee member once said after a brutal cycle: "I don't mind the work. I mind the guesswork."
The new math of admissions forces a choice: keep tolerating a system built on invisible sacrifice and unverifiable reasoning, or modernize toward clarity, equity, and resilience.
The first path keeps stealing our evenings. The second starts by admitting the obvious: the workload is already here. The only humane response is a workflow designed to handle it.
— Meshwell Staff
This essay is part of an ongoing series from the Meshwell policy team on rebuilding admissions with integrity. If your team is exploring auditable, outcome-linked workflows that reduce hidden labor, we’re happy to compare notes. Visit meshwell.ai to learn more and schedule a pilot.
