As AI reshapes the workforce, education must evolve just as quickly. In this episode, Jennie Sanders of Western Governors University, the nation’s leading competency-based, online university, shares how decision intelligence is being scaled to transform the student experience — enabling personalized learning, proactive engagement, and better decisions at scale. Now in use across thousands of learners, Jennie discusses what it takes to improve student success and graduation outcomes, advance equity, and rethink how humans and systems work together to serve the student of the future.
Transcript
Fred Laluyaux 0:07
Hello, and welcome to the Decision Intelligence Podcast, where we have real, candid conversations with thought leaders, practitioners, and tech experts about AI, decision intelligence, and the future of work. I'm your host, Fred Laluyaux, co-founder and CEO of Aera Technology. I'm excited today because we're joined by Jennie Sanders of the Western Governors University, the nation's largest nonprofit competency-based online university, and Jennie leads digital innovation initiatives focused on improving student success and outcomes through AI and decision intelligence. Jennie, welcome. So good to see you, and thank you for joining the podcast.
Jennie Sanders 0:53
Thank you so much for having me. It's great to be here again, Fred.
Fred Laluyaux 0:56
Absolutely, so, maybe we start with quick intros about you, what you do, and then we'll talk a little bit about WGU. Is that okay with you?
Jennie Sanders 1:07
Absolutely. Yeah, so I am a scientist turned educator turned, I guess, educational technology innovator at this point. I've been with WGU for 13 years. I started my, kind of, educational career, actually in biochemistry and biophysics, but I was really enamored of this idea that we could deliver education differently to meet the needs, especially of adult learners, those who, if they weren't able to get to college between the ages of 18 or 24, or if they did, but you know, life happened and they didn't complete. That we didn't just let all of that human potential, you know, fly to the side, but instead provide opportunities for them to do that. So, I went into online education, thinking this would be the future, you know, digital education really was very interesting to me. So, I found WGU, and haven't looked back for 13 years, and I've moved from a faculty role to leadership roles, and now I'm the Vice President of Product and Technology for the Academic Delivery Business Unit here, and very focused on that course and instructional experience, as well as assessment experiences, and how we can make those more personalized and more scalable with AI and digital and decision intelligence,
Fred Laluyaux 2:24
So, I’m such a big fan, I've talked about what you do and what WGU does in general, but talk a little bit about WGU as an organization, . . . how many students, how many faculty, it's massive, so I’d love for you to talk a little bit about the mission and what does it mean to have competency-based education, you know, it's super interesting.
Jennie Sanders 2:51
Yeah, absolutely. So Western Governors, it's been around now for, we just hit 28 years, so we're quite young for a university. We were founded in 1997 by 19 western state governors in the United States who were looking to really figure out how do we offer a more scalable and accessible form of higher education to support adult learners and our workforce, right? How do we develop our workforce, and so we've been private and nonprofit from the beginning, those states funded us just a little bit at the beginning, about a million dollars per state, and since then we've been self-sustaining off of tuition, and so when we say competency-based education, if you're not an education nerd, you may not know what that means, and what that means is that rather than measuring the amount of time that someone is in a classroom, which, if you think of a typical semester, it's usually 16 weeks or so in the United States, you know, students start a course and they are all in that course for 16 weeks, and then at the end they get a grade, right, and that grade represents their performance in that course, ostensibly it's what they've learned or what hopefully they can do, and the standards of those grades varies across even instructors in the same course at an institution, certainly across institutions, and they wanted something different. They wanted something that said, ‘What are actually the competencies?’ And if someone comes to us and they've already developed some of that competency, ‘How do we give them credit for that? And how do we help them move more quickly?’ So, at WGU, we don't measure seat time – we measure learning. We measure competency. And so if someone who's been in the field, let's say they've been, you know, they've owned a business for 10 years, they know a lot about running a business, and they come into our business degree, there may be many courses where they look at the competencies and think, I think I already know how to do that, so they can then go to the assessment portion, and they can prove that competency and move on. They might be in that course for one or two days, so they can accelerate, and that means that they move at the pace of their learning, and if they need less time, they take less time. If they need more time, they can take more time. If it's something that they really need to dig into a bit more. And that means that our average... Oh, sorry, go ahead,
Fred Laluyaux 5:03
No, I was gonna say, and then this way you accommodate people who need more time, have a couple of jobs, and are trying to get that degree, so they don't have the pressure of completing within a period of time, they can take the time that they need, I think it's a big part of what you enable, right?
Jennie Sanders 5:21
Yes, yes, it's much more flexible. We have monthly starts, not just a set semester schedule. And what we find is, you know, we have 200,000 students now, and around 4,000 faculty members, and we find that this really works quite well for many of our adult learners. Of course, more students are coming to us who are of a different profile. They're now 18 to 26, which is kind of our fastest growing demographic. They're noticing what we're doing, and they also need flexibility. Yeah...
Fred Laluyaux 5:54
That's new, right?
Jennie Sanders 5:55
It is new. It is new for us. So we're learning, and it's another reason why this decision intelligence capability is so important, because we want it to be personalized anyway with the students that we had before, but now we have a greater variety of educational background, understanding, goals, etc. And so, how do we personalize to a diversifying student body, and then do that at scale? So that's what we're trying to do. Our mission is to provide pathways to opportunity, and we do that with career-aligned, workforce-aligned programs and certificates and things that we can show - that they can show that can build these competencies, and then easily show to their employers, these are the things I can do, not just my GPA. This is what I can do.
Fred Laluyaux 6:42
So, we'll come back to what you do with the AI, obviously, because this is a DI podcast, and this is an incredible topic. But I'm just curious in reacting to what you said. So I always pictured WGU for adult learners, and you're saying there is this trend of younger learners, 18 to 26. Why is that, do you think?
Jennie Sanders 7:04
I think there's a couple of reasons. One, we started to see this really start to surge after the pandemic, and I think the pandemic gave a lot of folks exposure to online learning, not all of it wonderful, of course, but you know some of them thought this is pretty convenient. This is something I can do if I have a job, if I have children, you know, or if I'm taking care of parents, things like that. And also, it became even more important for many of them to avoid the debt that we hear a lot about with student loans, etc. We're a very low-cost option, and in that we have these sort of six month terms, and within that term, they pay a flat fee. It's essentially a subscription, and then they can finish as many courses as they can, do the learning and demonstrate the skills, and that means that it's a faster option for most, and it's a less expensive option, it's a more flexible option, and it's often aligned. Well, we only build programs, honestly, in areas that have high growth in the industry, so it's also something that's much more connected to the economic opportunity that they're looking for.
Fred Laluyaux 8:10
That is so cool. I'm not an expert in your field at all, but I have a sense that this is something we're going to see growing. It even has me thinking about my kids, who are looking at colleges right now. They might take a degree here and do some additional degrees online with WGU. I'll talk to them about this, I think it's a good idea. So let's dive into DI, you know, in this podcast we talk a lot about supply chain, talk a lot about operations and sales and other other functions, but you're applying decision intelligence in actually a very purist way, but in a very different use case to solve a very different problem. Can you talk about the problem that you are solving with DI, and then we'll talk about how we're doing it, you're doing it?
Jennie Sanders 9:02
Yeah, yeah! So, if you think about, so in supply chain, right, decision intelligence helps to better align supply and demand, right, in a timely way. And I would say, in education, in online education, if we think about the supply, that's the resources, the support, you know, things like that, that we have to provide to students, but the demand can be actually difficult to determine. You have the students who will self-advocate, no problem. Generally, they're the students who would do just fine without us in some ways, that they're good self-advocators, and so they take up a lot of that supply. But then there's the students who really need us, but maybe feel shy, or you know, don't feel like they sometimes, in cases you know, they've had such bad educational experiences that they are sort of just hesitant to even reach out, right? And so, in that case, we need to understand that demand and be able to connect that supply proactively without them needing to ask for it. And to do that in a way that's very personalized to the moment and the context that they're in, so that we get them at the right time, right? So we can support them before they experience a type of friction or have a moment of doubt where they decide to leave, that they decide they can't do this, because we really believe in the ability of humans to learn, and that with the right support and flexibility, just about everybody can be successful, so that's what we're using decision intelligence to do, is to really understand that student journey, and not just at an aggregate level, but at that personal level, and then determine where are these precise kind of surgical opportunities for intervention that can make the biggest difference to their learning, to their course completion, to their persistence and ultimately to their graduation.
Fred Laluyaux 10:44
So, this is so fascinating. And when you guys came to us with that use case, it's like, wow, I don't know if we can do that, but let's give it a shot, and I think we're able to help you, obviously. But it's literally demand sensing, you want to sense based on data and dynamic data, which student has the need for additional support, capture that in real time, and then deliver a set of actions, recommendations we like to call them, for the faculty to engage, and so on, and so forth. So, talk a little bit about how it works, the volume, and then what results you're experiencing, and when did you start the project? Give a sense of how long it's been running and what results you're getting.
Jennie Sanders 11:34
Yeah, I mean, we started our very first teeny little pilot in December of twenty four, and, (man, the years right now, time is strange) and so, about 18 months, not quite 18 months ago, and that started just with a few, a handful of instructors, a couple of courses, just to kind of test it out as a proof of concept, and we're now using that same use case across, I think we're now at close to 40 courses and several 100 instructors and 10s of 1000s of students, and what we've been seeing. So, should I explain the use case?
Fred Laluyaux 12:16
Yeah, please do!
Jennie Sanders 12:18
Okay. So, again we're trying to be quite surgical and very proactive in what we're doing here, so we noticed that for students, and honestly many of them are from that kind of younger cohort that comes to us with less educational experience in higher education, and we were noticing their first term that if they took their, you know, a course, but they had their first, what we call an objective assessment, that just means it's an exam, it's machine scored, and they don't pass. Many of them took that as a signal that ‘I don't belong here, I can't do this,’ and they would drop just right away. And we thought, "Oh, that's such a shame. We're a competence-based place. If you don't pass the assessment, you go back and keep learning, and then you come back, right? Like it's not intended to be all or nothing, because learning just doesn't work like that. And so we thought, okay, how can we be more proactive? And so we use decision intelligence to help us build some decision models that brought in information about how the student was engaging in courses; how are they doing on formative assessments; how are they engaging in other areas of the university; what other signals do we have about them to answer if they were to take the assessment today. Would they pass — like, what's the probability they would pass? And if they took it and didn't pass, what's the probability they would leave? And then we use that, plus a couple of other indicators, to then surface to instructors in those courses a recommendation that says, hey, this student, we think they could use some outreach to prepare for their assessment, and the instructors can accept the recommendation as is, they can modify it, or they can dismiss it. And what we've seen is that the students who have received those interventions, whether they were modified or just accepted as is, those students are seeing a seven percentage point higher assessment pass rate, and for those who are not familiar with education – changes that large at the scale that we were working is very hard to do, so it's quite a triumph. And then the course completion rates were around like four percentage points higher as well, so we saw really quite remarkable success with that.
Fred Laluyaux 14:25
And just to bounce off your point, if you can share it, otherwise, totally understand, but the success for that project, what was the threshold for you to declare success for that project in terms of increasing success rate?
Jennie Sanders 14:39
Yeah, I mean, we were hoping, honestly, just to get at least, I think we had, we committed to the board for a 2.5 percentage increase, and of course, completion rate, and we were like, okay, we think we can do that, right, and then we blew it out of the water, we almost doubled it in terms of what we were able to achieve. And the best part for me is if we also have seen consistency in the outcomes there, that over time, this wasn't just like right at the beginning, right? Right, like this has been quite consistent over time, and so that's been quite exciting. And then, of course, since then, we've launched another decision around early course engagement to help our students engage earlier on for higher success, and we're about to launch very soon one that is predicting if a student is going on a financial hold. We found that was another thing that could be pretty indicative of drop...
Fred Laluyaux 15:34
Of course!
Jennie Sanders 15:34
And how do we make it easier for them to prevent ever going on a hold? Just make the administrative piece of it completely easier, the decision making around it easier for both the student and then also for the back office. So this one's kind of two pronged, so we're expecting to see some efficiencies in the back office and a whole lot more retention for the students who may have got a financial hold.
Fred Laluyaux 15:55
So, instead of loading your ERP data, your supply chain, you're bringing in Aera, basically, the student digital journey. And that's what allows you to learn and predict. And that's what allows you to generate the recommendations to the faculty to engage at the right moment with the student, and then retain them, and get them excited, showing them a bit of attention and love that makes them want to, you know, stay longer and try for longer, which is phenomenal! When you deploy this across, you know, different populations of students, did you see anything that surprised you? Any differences? Any kind of learnings from again, having deployed that about a year ago now?
Jennie Sanders 16:46
Yeah...
Fred Laluyaux 16:47
Any differences, by student population, by type of curriculum? Anything noticeable?
Jennie Sanders 16:54
Yeah, that's a good question. So, I mean, truthfully, it was relatively even across the board in terms of some of those outcomes. We did see, though, that it had the largest positive impact on the students who were experiencing what we call momentum, like the lowest momentum, right? So some folks will call the students at risk, I don't like calling them that way, because they take a risk on us, right, so the students who were lower momentum, they actually tended to benefit more. We saw a larger delta for them, and so that was quite wonderful to see. I think the other thing that was interesting and has evolved, actually, over the last year was when we were looking at the results for the first maybe six, seven months, we saw that the accepted as-is recommendations kind of outperformed the modified recommendations, so if an instructor had, like, modified it, you know, that we still saw a lift, but it wasn't as large of a lift if they just took the recommendation.
Fred Laluyaux 17:54
Can you talk about.. sorry, I didn't want to cut you off. What are the types of recommendations that the faculty receive, what is the system recommending to them?
Jennie Sanders 18:05
Yeah, good question. So we recommend some type of outreach, and it can be a message, you know, it could be a text message, it could be an email message, it could be recommending, you know, a call, like recommending for the student to schedule a call, it could be sending them content that's specific to an area that it looks like they might be struggling with right and we do have AI generating those messages so that they're personalized to the student. We use some RAG to bring in some of that data as well to save the instructors time, but then the instructor can change that modality; they can edit the message, as you know as they wish. And we really wanted to make sure we were not removing that professional judgment, because there's personal relationships there — they will know things that, you know, the system doesn't necessarily know. So we wanted to make sure we kept the human in the loop on that one, but it's just interesting because I think over time, we're now seeing the modified outperform the recommended, and I think part of it's because the instructors are learning from the things that they're learning from, from Aera, essentially like from the decision intelligence piece, and so now they're making it, they're taking what they're learning there, and then they're adding even more of their professional judgment. Rather than sort of the default of what they would do, which is what was a bit more what was happening at the beginning, so it's been interesting to see the machine outperforming the human, and now the humans with the machine outperforming the machine alone, and so, that to me tells a really wonderful story about some of the things that can happen as we kind of use humans and machines to solve problems.
Fred Laluyaux 19:36
That's super interesting. I think I want to take this one offline with you, because I think there are some new capabilities in the platform that could be super helpful in learning what the faculty members are adding to the initial recommendation from the tool, so we're building a lot of cool stuff on that topic. Can you give a sense of the volume? You talked about, number of students, but how many recommendations? Are we talking, are we talking thousands? I know the answer, but I think it's good for you to share it!
Jennie Sanders 20:08
Yeah, we're over, I think we're about 8,000 recommendations that have been actioned. How many have been surfaced? I think over 15,000, have been surfaced. And so, yeah, about 8,000 recommendations, actioned recommendations at this point.
Fred Laluyaux 20:25
And yeah, that's a lot. And if you look at the evolution – so you went live, now you're talking about the evolution of going from the recommendations to the modified recommendations – what other learnings did you get in the last year since you went live? Anything else that you want to share, beside that point, which I think is super interesting.
Jennie Sanders 20:51
Yeah, well, I mean, I guess it kind of confirmed the hypothesis for us that moving upstream, being more proactive, can have a really large impact on the student experience and their success. There's also with our faculty, the other thing that we learned, and in some ways it's relearning, right, is how that's recommended to them and the amount of agency that they feel around it makes a really big difference to adoption, and then making sure that it's embedded within an experience for the instructors that helps them take that recommendation with some additional context, and so we've been working on that in terms of the UI for them as well.
Fred Laluyaux 21:33
So, we talk a lot about the impact. Ultimately, the impact is for the students, you want to retain them, you want to encourage them, you want to give them what they need, that little push, that little content, or sometimes that engagement, that simply may be what's needed. But what about the instructors? What about the faculty? Is this more work, less work? How did you get adoption? Were they're like, "Hey, what the heck, I'm already so busy, I can't do everything”, or they say, "No, actually, this is helping me,” Just completely transparently, how did you drive the adoption on the faculty side, on the instructor side?
Jennie Sanders 22:10
Yeah, I mean, luckily at WGU we have a lot of faculty who love to innovate, they love to try new things, and so when we were running, rolling this out, we had a number of them that were actually quite eager to adopt, and then as they did that, their colleagues started to hear about it, they wanted to try it too. Yeah, and it starts to get better and better. And actually, we presented at SXSW EDU this past March...
Fred Laluyaux 22:36
Nice!
Jennie Sanders 22:36
Myself, my product manager, and we actually brought one of the instructors with us — Mijung Kim is her name. And she joined us, and she shared her experience using it, and one of the things that she described was before she would come in, and it would take her quite a bit of time to prioritize, who do I reach out to right now and are they the right person, et cetera. And she loved that with the decision intelligence recommendations, she could go there first, see the things that would be like highest priority, right. And then she could still use her professional judgment on exactly how to word something for a particular student. But it saved her a lot of time!
Fred Laluyaux 23:12
Yeah...
Jennie Sanders 23:12
So, she said it gave her a lot of confidence. You know, it streamlined her workflow. And, it's not often that faculty want to... How do I say this delicately?...
Fred Laluyaux 23:28
It's only us right now!
Jennie Sanders 23:29
Yeah, yeah, faculty are experts, right? They have PhDs, they spend a lot of time studying things, and so they like to be right all of the time.
Fred Laluyaux 23:37
Yeah...
Jennie Sanders 23:38
But it was interesting to hear her say, "Oh, I never would have actually selected this student, and it made a really big difference for them, and so they were learning from this experience about how to hone their craft, and that was really, I mean, it was great to hear Mijung talk about it, she was one of several, of course, that have said that they love it, they actually just want more recommendations, she asks me on the regular (basis), she'll IM me and say something like, "So when are we getting more recommendations?" And I’m like we’re building them, we're working on that!
Fred Laluyaux 24:09
It just makes me so happy to listen to that, because I think what you guys do — and I've told you before — is just absolutely amazing. I really embrace the vision of WGU and what we're doing, but how you're leveraging technology to solve and ultimately get more performance from the system that allows people to get educated who otherwise would not have, potentially have, access to that education. I mean, this is life-changing stuff, right? For the people who are following your courses. So super proud to be partnering with you. What's next? So, you talked about that faculty member asking for more recommendations. What are you working on right now? What is Jennie's hat?
Jennie Sanders 24:47
So, there's a number of things, so we're looking at high-impact moments across that student journey. So, I know we talked about the financial hold one. We actually have another one that is in beta right now in our enrollment funnel to help our enrollment counselors again prioritize who they're reaching out to, students who might be feeling stuck on a decision, things like that, but then there's several others that we're working through and building out, so one is how to detect when a student might be engaging in a way that's different for them that signals what we're calling silent drift. Often students, especially in an online environment, it's not like they, you know, slam the door and leave, they just sort of silently drift away. And there are times when they might drift a little bit because they're on vacation or they're having a busy time at work or something's going on at home, and so this is now where we're thinking about how do we really understand for this student what's normal in terms of that engagement pattern, what's not so normal, and then what's the right intervention for them at the right time. And that's something that is, I was joking with the product manager about this, I'm like, if you solve this, you've really nailed a really critical problem in online higher education, because that's a tough one, like really understanding the type of engagement pattern that works for different types of students.
Fred Laluyaux 26:07
So, yeah, that's getting the data and deploying the right algorithm and fine tuning the algorithm to be able to detect this stuff. Well, that's exciting, Jennie, any thoughts for the folks in your industry who are listening to us right now? Any recommendations that you want to share?
Jennie Sanders 26:33
Yeah, I think for me some of the unlocks with decision intelligence that I think could be used across education is that first we're starting with the end in mind with that outcome you know that we're really seeking, and then through this technology and through these systems being able to develop causal models that help us get to that outcome so much faster than some of the more generic algorithms or heuristics that we've often used in the past, so I think that the decision intelligence methodology is incredibly powerful, and I think humans are in some ways a harder use case than supply chain, because you know, in supply chain, your products don't get to make a decision about whether they get on the truck or not, so you know we have that to contend with, but that's also what these educators are so good at is working with these wonderful humans, and giving them the support that they need, and the students, some of these direct to student support that they need in a timely way that is truly personalized. I think it is just game changing for what we can offer in terms of that personalized learning and support throughout that journey.
Fred Laluyaux 27:39
Jennie, this is so exciting. I could speak with you for hours, like always. I'm so proud of the partnership, one of the ones that makes me really happy, both personally and professionally. Very impressed with what you do, and the success that you've had. Thank you for the partnership. Thank you for sharing your thoughts today. And, as always, we're, you know, one phone call away, or email, to help you, and I still need to come and visit...
Jennie Sanders 28:05
You do!
Fred Laluyaux 28:05
...graduation. I know, I'm gonna... I told you I'm gonna do it. I'm a man of my word! I will do it, I just need to find time, but...
Jennie Sanders 28:11
We have one coming to Anaheim pretty soon, so...
Fred Laluyaux 28:13
Okay, that might be the one! Let me check! But I really want to come and see. You told me about those ceremonies being so emotional, so that's something I don't want to miss. Thank you so much for joining us today, Jennie. And thank you for the great work.
Jennie Sanders 28:28
My pleasure, Fred. Thank you for having me.
Fred Laluyaux 28:30
All right, so this concludes our podcast, Decision Intelligence in Action: Serving the Student of the Future. You can view today's podcast on our website, www.aeratechnology.com. Thank you very much. See you next time.