The Upskilling Trap: Why Domain Experts Are the Real AI Advantage

In higher ed, the gap between AI potential and AI reality isn’t a technology problem. It’s a people problem.

A while back I started following Hilary Gridley, a former product leader at Whoop who’s been writing about AI adoption and use. Her (first) episode on the How I AI podcast was a game-changer for how I started using AI differently. I’ve been sitting with one of her recent Substack articles and had to share it because it hit so close to home.

Her argument is that the people best positioned to drive real AI transformation are those rare individuals who have both deep domain expertise and technical skill. Not one or the other. Both. And she’s right that they’re hard to find.

I’ve spent two decades in California community colleges moving between worlds that don’t usually overlap. Counseling and tutoring students with disabilities. Teaching remedial math and English. Troubleshooting and optimizing Ellucian Colleague and related systems. Managing federal grants and navigating Title IV compliance. Writing reports. Leading a team through financial aid processing while chairing a Technology Committee.

I’ve spent my career thinking of this as an unusual winding path, and occasionally wondering if I should have instead specialized in doing one thing really well. Reading Hilary’s piece, I’m starting to see it differently.

The domain expertise in higher education is deeply specialized. There are federal student aid regulations, state mandates, strategic enrollment management, equity-centered decision-making, the lived realities of students navigating poverty and disability and bureaucracy all at once. It takes years to develop (especially if you’re working in California). You can’t consultant your way into it.

Hilary’s framework surfaces something I’d add for the higher ed context. The stakes for getting the human piece wrong are different here. When community college employees feel sidelined or replaced, the people who lose aren’t shareholders. They’re first-generation students who needed someone in the room who understood their situation.

The AI work I’m most proud of at my institution has happened at exactly this intersection: compliance knowledge meeting technical capability, with students at the center. That’s not something you can hand off.

If you’re in higher education and thinking about where AI fits in your work, I’d highly recommend Hilary’s piece.