Sentara just put AI to work in case management — what it means for healthcare admins

Dr. Ashley Hussain-Okorafor, DBA
Former lecturer and hiring-side healthcare administrator

Let me be direct about what just happened: a hospital system put an AI tool to work in a job most patients never see — and it's the clearest signal I've seen this year about where healthcare administration is actually heading.

Last week, Sentara Health went live with an AI case management tool it co-developed with a clinical decision support company called Regard. Sentara is a not-for-profit system running 12 hospitals and more than 400 care sites across Virginia and northeast North Carolina. Case managers can now type plain-language questions and instantly pull up a patient's history, status changes, and what's shifted in the last 24 hours — instead of digging through a chart by hand. Sentara's chief health information officer put it plainly: the team used to spend all that time "piecing together a patient's story" before discharge planning could even begin.

Here's the number that should get your attention. Sentara executives estimate the tool could save nearly 90,000 hours — the equivalent of adding more than 40 full-time employees and close to $5 million in labor costs.

Let me sit with that for a second, because it's the whole ballgame.

This is AI leaving the bedside and walking into operations

Most of the AI stories you've heard in healthcare are clinical: a model flags a diagnosis, reads a scan, or suggests a medication. Those matter, but they live in a world administrators visit, not one they run. Case management is different. It is operations. It is discharge planning, length of stay, throughput, denials, and getting a patient safely out of a bed so the next one can use it. That's the exact territory a healthcare administrator owns.

When AI starts doing that work, it stops being a doctor's tool and becomes an administrator's problem — and an administrator's opportunity.

What case management actually is (in plain English)

If you're new to this, here's the honest explanation. When a patient is admitted, someone has to figure out where they go next: home with services, a skilled nursing facility, or nowhere until a barrier clears. That someone is the case manager, and their decisions touch almost every metric a hospital is graded on — length of stay, readmission rates, and whether the hospital actually gets paid for the care it already delivered. The work is reading charts, coordinating, documenting, and negotiating, day after day. A lot of it is manual. That's the part this tool eats.

Notice what the tool is not doing: it isn't deciding where the patient goes. It's pulling the story together faster so the human case manager can decide. That distinction is the difference between "AI is coming for your job" and "AI is coming for the boring half of your job." Most of what I've seen says it's the second one — and that's good news for you, but only if you understand the game.

Why this changes what gets you hired

Let me translate this into your career, because that's the part most news sites skip.

Ten years ago, a healthcare admin could get by knowing their department and their spreadsheet. The person who gets promoted over the next decade will be the one who can sit between three groups that don't speak each other's language: the clinicians who document, the people who build the tools, and the finance team who count the hours and dollars. This tool sits exactly in that intersection.

Here's what I'd tell you to do with this story:

• Learn the operational vocabulary. Length of stay, throughput, denials, days in A/R, readmission rate. When you can talk about 90,000 saved hours as hours of discharge-planning capacity — not "AI magic" — you sound like the person who manages the tool, not the person waiting to see what it does.
• Get comfortable translating. The case manager knows the patient, the analyst knows the model, the CFO knows the margin. The admin is the one who turns "this tool surfaces patient history faster" into "this reduces avoidable days and denial rework." That translation is a skill, and it's learnable.
• Treat AI as something you govern, not something you fear. Every one of these deployments needs a human who decides where the tool belongs, what it can and can't do, and how to tell when it's wrong. That's a leadership job with your name on it if you want it.

The honest caveat

I'm not going to sell you the hype. "Could save nearly 90,000 hours" is an estimate from the people deploying the tool, and estimates have a way of shrinking when the bill comes due. The case managers aren't being eliminated — Sentara is framing this as giving hours back, and I'd watch whether those hours become faster discharges, more patients served, or something quieter. The workforce question is real, and I'd rather you walk into it with your eyes open than get blindsided by a headline later.

But here's the thing I keep coming back to: whichever way those hours shake out, the person who understands what they're worth is the person who keeps their seat at the table. That's the healthcare administrator.

What I'd do if I were starting today

If I were in your shoes — student, career-changer, or early-career admin — I'd do three things this week. First, look up one of those metrics I named and learn what actually drives it. Second, reread this story and write down every word that belongs to operations, not medicine. Third, ask yourself whether you'd rather be the person deploying the tool or the person whose job it was pointed at. You want to be the first one, and that's a choice you can make starting now.

AI didn't come for healthcare administration this week. It came for the busywork. The people who thrive are going to be the ones who know the difference.