The New Normal, Episode 7: Dr. Greg Fowler

Five takeaways from the president of a global university serving more than 100,000 students on building from the problem backward, giving military rank academic credit, and forecasting where students will need help.

University of Maryland Global Campus traces back to 1947, when seven faculty flew to Heidelberg to educate service members stationed in Europe. Today it serves a little over 100,000 students across 20 to 24 countries.

Greg Fowler is UMGC’s seventh president, and the seventh guest on The New Normal, OutRival’s interview series with leaders deploying AI in production.

Here are five takeaways from his conversation with OutRival co-founders Ruben Harris and Timur Meyster.

1. The problem comes before the technology

Fowler borrows an analogy from Michael Horn that he heard at a recent legislative hearing.

When electricity arrived, the early productivity gains were disappointing because factories used it to do what they had always done. The gains only showed up after people rethought the work itself.

UMGC’s version of this idea is refusing to bolt AI onto systems that were not built for it. Customize your way out of a bad fit, Fowler says, and you tend to break the thing you were building or slow it down.

Instead, the current investment is in the unglamorous foundation: data lakes, warehouses, and platforms, plus business analysts who can map how work actually moves through the institution.

The payoff of understanding a workflow is that sometimes the answer turns out to be a better workflow. Sometimes it is automation that does not require AI at all. “We can’t just be chasing headlines,” Fowler says.

2. Military rank for credit

The clearest example of that problem-first approach started with a conversation in Fuji, Japan.

Fowler met a sergeant major about to graduate who mentioned that his early coursework had included a 200-level class in organizational leadership. Fowler pointed out that as a sergeant major, he had likely been leading organizations for years. The sergeant major agreed he probably could have taught the class himself.

Rank progression already contains the three things a course requires: an assessment, an expert qualified to evaluate it, and a documented process. The obstacle was that military skill sets change fast as service members move up in rank, and older credentialing frameworks like ACE credit couldn’t keep up.

UMGC used AI to match the skills demonstrated at each rank level, from E4 up through E7, against the skills its courses already teach. The program launched last October, and so far, more than 20,000 service members have used it, collectively saving more than $20 million in tuition. Fowler expects first responders to be the next population evaluated the same way.

3. Forecast the student instead of reacting to the transcript

UMGC applies AI to the student experience itself through two projects under its Strategic 2030 program.

The first is Project Sherpa, a personalized assistant available at all hours, which matters at a global institution where someone is always awake somewhere in the world.

A Sherpa cannot climb the mountain for you, Fowler says, but climbs it with you, and notices which climber needs to slow down and which one is showing early signs of altitude sickness. In practice, that means a system that learns a student is afraid of math during enrollment, then reaches out when the math requirement arrives months later.

The second is Project Doppler. Fowler compares the way it works to meteorology, pulling in many factors at once to predict what’s ahead. Just as weather forecasting combines temperature, pressure, and wind to predict a storm, Doppler combines a student’s background, such as a veteran from Missouri entering a particular field of study, to predict where that student is likely to struggle.

Subscribe to The New Normal here.

4. Move the effort from leads to applicants

Asked how UMGC thinks about the enrollment funnel, Fowler talks about how they’ve shifted their emphasis. His team spends less attention on the inquiry and lead pool and more on the applicant pool, because leads that never convert do not help anyone.

Once someone applies, UMGC still needs to know who they are. Applicants now include not just military students but also, for instance, retail employees, financial advisors, and defense contractors, and each group tends to struggle in different ways. That is where the forecasting from Project Doppler comes back in, this time for the application stage.

5. Chatbot, copilot, or agent

Fowler puts UMGC’s tools into three tiers based on how much judgment they carry.

A chatbot is the basic bubble at the bottom of a website. It answers simple questions with no real judgment behind it.

A copilot, like Military Rank for Credit, works alongside a person who still makes the final call.

An agent, meanwhile, acts and learns on its own. For example, UMGC’s cybersecurity systems watch for phishing attempts every day and learn from what they see, getting better at catching them without a person needing to review each event.

On whether to build these tools or buy them, Fowler compares it to assembling a large jigsaw puzzle. No one is going to sell you the finished picture, so you start with the pieces you can place with confidence, like flat edges or a matching color, and build outward from there as more pieces become usable.

Listen to the episode

Check out the full episode to hear:

• Why Fowler thinks the future of credentials looks more like a playlist than a compact disc
• What the Alpha School model gets right about knowledge and social interaction, and what he wishes it emphasized more
• The 85-year-old graduate whose classmate rode the bus with her when he was a baby on her knee
• The Yevtushenko poem Fowler returns to when he watches thousands of students cross the stage
• Why students who cannot go to Maryland choose Penn State or Michigan rather than the next school down the list