OutRival Presents: The New Normal

OutRival CEO Ruben Harris sits down with Lisa Marsh Ryerson, President of Southern New Hampshire University, to break down how one of America's largest universities serves 300,000 students with AI agents working alongside human advisors.

SNHU invented Speed to Lead, the practice of reaching students the moment they raise their hand. Now the university is building on that foundation with an AI native strategy: agents that handle the hand work so advisors can do the heart work, warm handoffs from chatbot to human, and an executive team the president personally put through a six week AI boot camp.

President Ryerson also shares why 300,000 learners is no longer scale in her mind, how SNHU evaluates AI vendors as strategic partners, why she tells her team they cannot tweak their way to the future, and what it takes to aim for 10 million learners.

The New Normal goes inside organizations using AI in production at scale. No hype. Just leaders who reimagined their organizations with AI at the core.

Key ideas from the conversation:

• Speed to Lead exists because the moment a student raises their hand is fragile. Life rushes back the next day.
• If you only answer inbound, you are totally reactive. The work requires being daily, intentionally proactive.
• AI agents give advisors an Iron Man suit: tools that surface who needs attention today.
• The plan should never be to buy a bunch of licenses first. Understand your data structure and your people's capacity to implement with fidelity and ROI.
• Risk aversion and risk taking coexist through a disciplined process: name the risk, mitigate it, manage the residual, and write after action reports.
• You cannot tweak your way to the future. Systems that are not AI native require whiteboarding from what is true about learning.
• Higher ed should never cede authority over what learning is to the frontier labs.
• Great vendors are not vendors. They are strategic partners who embed with your teams and obsess about your learners as much as you do.
• Personalization is not static, because goals and purpose change across a lifetime.
• Hand work belongs to agents. Heart work belongs to people.
• 300,000 learners is no longer scale. The challenge on the table is 10 million.

Learn more about OutRival: https://www.outrival.com 
Watch every episode: https://thenewnormal.fm

Follow Ruben Harris: 
LinkedIn: https://www.linkedin.com/in/rubenharris 
Instagram: https://www.instagram.com/rubenharris

Follow The New Normal for new episodes every Thursday.

#AI #HigherEducation #SNHU #AIAgents #EdTech

Rubén Harris sat down with Dr. Mark Milliron, President of National University, on campus in San Diego, to talk about running a university where the competition isn't other institutions. It's poverty. Mark grew up in a family of nine kids with 25 foster kids rotating through the house, bused tables to pay his way through Mesa Community College, earned two degrees at Arizona State and a doctorate at UT Austin, then built a career that spans the Bill & Melinda Gates Foundation, an early board seat at Western Governors University, corporate work at Oracle, and the chairmanship of the Trellis Foundation. Today he leads National University, which serves 130,000 learners a year. Almost all of them are working students, half are parents, half of undergrads are military-affiliated, and the average undergrad is 32. National calls them Anders: students and parents, students and employees, students and deployed. If you ask them to choose between their ands, you lose.

We get into the Wingman initiative that aims to put an AI Agent next to every employee and student, why grad students were the original AI agents, the four rings that replace a chief AI officer, the RAISE 5 initiative, the framework that codes every assignment for how much AI students can use, and the story of a faculty member who threw his computer out a second-story window and later opened his email for the first time to save one student's degree.

Learn more about National University here: https://www.nu.edu/

𝐒𝐇𝐎𝐖 𝐍𝐎𝐓𝐄𝐒 (key moments and quotes)

Competition is poverty. "My competitors aren't institutions. My competition's poverty." Higher ed shares ideas, which Mark jokingly calls the CASE method: copy and steal everything. His partners in common cause include WGU, UMGC, and Miami Dade through the Presidents' Forum.

Anders, not students. National's learners are students and parents, students and employees, students and deployed. "If you ask them to choose between their ands, you're gonna lose." Designing for Anders is a feature, not a bug.

Put AI on purpose. The worst thing to do with AI is "we're going to do AI." AI is not a business tool. It's universal infrastructure, and it only works if it's attached to mission.

The Wingman initiative. Every employee and every student gets an AI wingman. The insight came from the Gates Foundation: most students who leave do so in good academic standing, because life happened. The fix isn't the curriculum. It's navigation. Even 30-year veterans of NU discover resources they never knew existed, and no single human can hold them all. A wingman can.

Grad students were the original AI agents. Every graduate professor in America has been using agents for hundreds of years. They were called grad students. Now it's grad students plus AI, and the game moves toward collective intelligence.

Use student data for the student. Institutions use data to prove their worth to accreditors. Mark's position: process it in near real time to help the student make a better choice and go in a better direction, so personalized education isn't reserved for people with half a million dollars.

Get grounded, get real, get going. Then try, test, tune. Every bet gets instrumented. Bad data can't trigger a culture of blame. "A culture of blame will kill you. It's gotta be a culture of wonder."

RAISE5. The Rules for AI in Scholastic Environments: five levels, from no AI at all to deep agentic collaboration, coded onto every assignment. When the law school rolled it out, 95% of assignments were level five, no AI. Then the conversation started.

No chief AI officer. A whole-of-institution AI council built on four rings: student experience at the center, academics and operations, infrastructure, and governance. The leaders over all of it are jokingly called the Lord of the Rings. More than 150 people now lead AI initiatives across NU.

Steve and Gina. A tenured faculty member threw his newly installed computer out a second-story window. Years later he knocked on Mark's door, closed the blinds, and asked to have his email opened for the first time, because a student fleeing domestic violence had emailed him her final assignments. She walked across the stage as a nurse.

Novel to normal. Email, microcomputers, the internet, and now AI. Every technology follows the same curve, and Mark has seen this play before. Innovators need data, the early majority needs evidence, and the late majority won't move without playbooks.

Educational malpractice. Not teaching students the AI of their discipline is putting your head in the sand. So is not teaching durable skills. Both are failing to prepare people for the world as it is.

The market nobody sees. Less than 20% of higher ed is the full-time, on-campus, non-working student. 43 million Americans have some college and no credential, and the vast majority left in good academic standing. "If we'll take 1%, you can have the rest."

The 280Z. Mark's dream car couldn't fit his 6'4" frame, and the lesson became his philosophy of fit: there are students for whom traditional education is perfect, and there are Anders who need a different model. Better with, not better than.

Rubén Harris and Timur Meyster sat down with Dr. David Andrews, Chancellor of UMass Global, at the university's Southern California headquarters, to talk about running a fully online university for working adults in the AI era. David wrote the first dissertation at Florida State University on a word processor, and his committee accused him of cheating because it wasn't a typewriter. He went on to serve as Dean of the School of Education at Johns Hopkins, where the school ranked number one in the country two years running, then as President of National University. Today he leads UMass Global, serving 19,000 students across all 50 states and eight countries, 97% of them working adults with an average age of 35.

We get into why roughly 60% of UMass Global's efforts now involve AI, why faculty are trading grading for rubric-building, and the belief David gets the most pushback on: education should be easier.

Learn more about UMass Global here: https://www.umassglobal.edu/

𝐒𝐇𝐎𝐖 𝐍𝐎𝐓𝐄𝐒 (key moments and quotes)

Accused of cheating, in 1980. David wrote the first dissertation at FSU on a Xerox 820 with eight-inch floppy disks. His committee said it was cheating "because you can change the words too quickly." Every first mover gets accused of skirting the rules.

Organized around people, not students. "A lot of higher education is not organized to take advantage of the technology that exists. In fact, we're organized around people." The delivery model is more faculty-centric than the technology requires.

Reactive vs proactive. Reactive AI is now ubiquitous; you have to opt out. The proactive side, reaching out, understanding who students are, and predicting what they need, is where most universities still aren't participating. Roughly 60% of UMass Global's efforts are engaged with or exploring AI, from marketing and enrollment to scheduling and the student experience.

From 48 hours to 48 seconds. Paper grading turnaround used to be 48 hours. With detailed rubrics applied by AI, it happens in seconds, and faculty now spend their time building better rubrics instead of grading one paper at a time. As Timur points out, professors are becoming prompt engineers whether they realize it or not.

Serial monogamy with instructors. What one professor learns about a student almost never transfers to the next. "If you're not collecting information at a micro level, then all the AI in the world is not gonna help you personalize."

50 commencement speeches, all written with AI. Last year, eight submissions. This year, 50, every one AI-assisted, including the winner, an orphan raised in foster homes until 19. "Why would I diminish her experience because it was assisted?"

The heresy. Where does David want to deploy next? Fully student-facing AI without human intervention. "Are we waiting for the technology? No. We're waiting for the fear to reduce."

Pilots, then pain points. Start with the coalition of the willing, then find the problem that touches 75 to 80% of the institution. "You don't fail, you pivot."

Sorting vs serving. "The expectation is that we're supposed to sort people, not serve people." And the belief he gets the most visceral pushback on: faster, cheaper, and yes, easier. "We're trying to make everything else in life easier. What's wrong with that?"

Rubén Harris and Timur Meyster sat down with President Scott Pulsipher of Western Governors University, at WGU's headquarters in Salt Lake City, to talk about running the largest nonprofit online university in America in the AI era. Scott didn't climb the academic ladder. He spent more than 20 years building technology companies, including a run at Amazon, before taking the presidency in 2016. Today WGU serves nearly 380,000 individuals a year across all 50 states, carries roughly 190,000 concurrent students, and has awarded more than half a million degrees since its first graduate in 2000.

We get into the Amazon lesson he carried over (customer obsession becomes student obsession), why WGU calls itself a student-centered, tech-first organization, the transcript broker service that pulls your prior credits for you, what actually separates an AI agent from a copilot, why the human has to stay the principal, the mentor math that could take one mentor from 100 students to 200, the “replication trap” that kills most online programs, and the unit economics behind a sub-$20,000 bachelor's degree.

Learn more about Western Governors University here: https://www.wgu.edu/

𝐒𝐇𝐎𝐖 𝐍𝐎𝐓𝐄𝐒 (key moments and quotes)

• Scale, one student at a time. Nearly 380,000 individuals served last academic year, roughly 190,000 concurrent across all 50 states, nearly 70,000 graduates this year, and more than half a million degrees awarded since 2000. WGU is 30 years young.
• A different definition of quality. Conventional higher ed measures quality by how many it turns away. “Our quality's gonna be measured by how many lives were changed by education.” Opportunity doesn't shrink when more people participate in it. It expands.
• Student-centered, tech-first. Just over 10,000 employees, nearly two-thirds of them interacting with students daily. A traditional institution serving 50,000 students might carry 25,000 to 30,000 staff. “That doesn't happen without a heavy emphasis on technology.”
• The Amazon transfer. Customer obsession becomes student obsession. NPS in the 70 to 75 range, roughly double the higher ed average and in Apple territory. 100% of employers in the most recent survey hired a WGU graduate again.
• “It's a process problem.” The Amazon default when outcomes vary from intent: don't blame the people, fix the process. Employees come and go. The process compounds. It's how WGU went from 65,000 concurrent students to nearly 200,000 in a decade.
• Automate the simple, agent the complex. A transcript broker service pulls prior credits on the student's behalf. AI maps course learning outcomes where no articulation agreement exists. Agentic AI carries the advising conversation.
• What an agent actually is. “An agent is someone who has authority to act on behalf of a principal.” Name the principal first (the enrollment counselor, the mentor, the course instructor), then design the agent to reason the way that principal would.
• Agents that work for the student. Ingest the lecture content, synthesize the learning outcomes, pull third-party resources, build the study plan. Spin up five and you have an AI-powered study group of one human and four agents.
• The mentor math. Roughly 80 to 100 students per mentor today, and they don't all need the same engagement. Push the lower-value touches to agentic support and one mentor could carry 200.
• AI on assignments is savvy. AI on assessment is not. “Then you're just assessing the copilot.” The fix isn't detection software, it's continual assessment instead of a midterm and a final.
• The replication trap. Don't move your lecture online. Rethink the design, the segment, and the cost model. “What about online is on a schedule today? I stream what I want to on Netflix when I want to.”
• Unit economics. The unit is the student term. Two-thirds of cost is direct delivery, 10 to 15% is technology, and there are no dorms or dining halls in the model. Students finish in about two and a half years for under $20,000. More than half are Pell eligible.
• Where the model came from. Governor Romer's pilot test: he didn't care where the pilot went to school, only whether they were competent to fly the plane. That's competency-based education, where time varies and mastery doesn't.
• Timing beats messaging. When a student fails an assessment, fast outreach from a mentor drops the second-attempt non-attempt rate to single digits. The modality shifted to text and chat. The outreach still matters.
• The dinner table. Six kids, four grandchildren with a fifth on the way, and the line that closes the episode: they cannot mistake their way out of your love. “Every single one of you is needed.”
• Eleven quarterbacks. “The worst team in the NFL would be the one with 11 quarterbacks on offense.” But higher ed talks as if every graduate is supposed to be one.

Rubén Harris and Timur Meyster sat down with President Madeline Pumariega of Miami Dade College, inside the college's own AI Center on the Wolfson Campus, to talk about running the largest community college in America in the AI era. Madeline arrived at Miami Dade as a student athlete, rose to chancellor of the Florida College System, and returned in 2021 as president. Today the college serves 125,000 students a year across eight campuses, counts 2.5 million alumni, and has trained 10,000 people and 400+ companies in AI.

We get into why she's “not a pilot person” and takes everything straight to scale, cutting a 45-minute admissions application down to 7 minutes, the $10,000 faculty AI grants now in their third year, launching Florida's first AI degrees, the difference between digital workers and AI agents (and why digital workers belong on the org chart), why students engaged longer with digital workers than with humans, and what it means to lead a 20th-century institution delivering 21st-century talent.

EPISODE LINKS:

Substack: https://outrivalpresentsthenewnormal.substack.com/

Follow us on Twitter: twitter.com/outrivalAI

Follow us on Instagram: instagram.com/outrivalAI

Follow Madeline Pumariega on LinkedIn: https://www.linkedin.com/in/madelinepumariega

Learn more about Miami Dade College: https://www.mdc.edu/

Rubén Harris and Timur Meyster chatted with our friend Kermit Cook, CEO of Penn Foster, about reinventing a 135-year-old distance learning company for the AI era. Penn Foster was founded in 1890 to teach coal miners in Scranton, Pennsylvania. Today it graduates 93,000+ learners a year and is on a path to graduate 150,000 by 2029.

We get into deploying Digital Workers (AI Agents) to support working adults at $1,500 high school tuition, why Penn Foster is the third-largest "school district" in the country, the unusual journey from Dartmouth engineering to KKR to Penn Foster CEO, how to stop treating AI as a tech project and start driving real outcomes on enrollment, retention, and ROI, and what it actually takes to modernize an institution that has served learners through every major technological shift since the telegraph.

Learn more about Penn Foster here: https://www.pennfoster.edu/

𝐄𝐏𝐈𝐒𝐎𝐃𝐄 𝐋𝐈𝐍𝐊𝐒

• Listen on Apple Podcasts: https://podcasts.apple.com/us/podcast/outrival-presents-the-new-normal-ep-1-penn-foster-ceo/id6783387190?i=1000773897953
• Listen on  Spotify: https://open.spotify.com/episode/3n1cTJ4mExmCeMd8teulVc?si=ZPNFgZfyQIe9M83ZaHcptw
• Listen on Soundcloud: https://soundcloud.com/newnormalpodcast/episode-1
• Subscribe to our Substack: https://outrivalpresentsthenewnormal.substack.com/
• Follow us on Twitter: twitter.com/outrivalAI
• Follow us on Instagram: instagram.com/outrivalAI
• Follow Kermit Cook on LinkedIn: https://www.linkedin.com/in/kermitjcook/