D2L CEO John Baker: “We’re Leveraging AI to Make the Educational Experience More Human.”

In 1999, a student at the University of Waterloo (based in Waterloo, Ontario) by the name of John Baker founded a company with the mission of improving the way that people learn. Two decades later, D2L, which offers its signature LMS Brightspace, has offices and institutional partners around the world. The company survived the dotcom crash and a three-year patent violation lawsuit brought by Blackboard in 2006. Now, in 2019, D2L has the third largest market share in the North American higher education market in terms of enrolled users, according to eLiterate. Since that analysis was published, D2L has signed some significant new clients, like Purdue University and the Maine Community College system.

Like many LMS providers, D2L is operating in the K-12, higher education, and corporate sector in markets around the world. Besides North America, the company also has employees in the Netherlands, the U.K., Singapore, Australia, Brazil, and Colombia. Between their 13 different offices, they employ a team of roughly 800 professionals and have clients in nearly 40 different countries.

“So we have about 1,500 total clients,” D2L CEO John Baker said over the phone. “Millions of folks login using our platform every single day to learn, both in the education and corporate sector. Our roots are in the education space, but more recently, we’ve been picking up a lot of big corporate clients as they tackle the re-education, reskilling, and upskilling of their workforces to support the future of work.”

But while most other LMS providers place a greater emphasis on one sector over another, Baker says that D2L’s services are more evenly spread.

John Baker’s remarks below have been edited from the original interview transcript.

John Baker: In our case, we’ve got growth in every segment. We’re the market leader when it comes to large virtual schools across North America and many places around the world where online learning is a priority. We’re also making great progress in terms of supporting millions of kids in K-12 schools with blended learning products. 

We’re also continuing to skyrocket in terms of number of users in the universities and college sector. We’re seeing dozens and dozens of new clients coming on board in that space. And we’re seeing real momentum shift towards not only the learning experiences of today, but also to help support the transition to new models of learning, like competency-based education.

Then in corporate, it’s a red-hot market. We’re adding lots of great clients, everyone from folks like Accenture or PNC Bank, Amex, you name it, there’s lots of large companies that are really trying to upskill and reskill their workforces right now. We’re at the center of that.

D2L’s Brightspace Acts as a Core LMS that Can Be Used by Any Company or Institution in Any Region and in Any Sector

Most LMS providers tailor their product for the markets they serve. Most LMSs, furthermore, tend to see most of their growth in individuals sectors. But D2L’s Brightspace is more evenly spread. That affects how the company develops it. D2L treats Brightspace as a core offering that can be applied, tweaked, and adjusted for just about any learning environment.

John Baker: In our case, we’re kind of unique, we have one learning platform, we build features for different segments. But in 90% of the cases, the features that we’re building for one segment have a halo effect, if you will, improving the outcomes or improving the experience for other segments of the market.

So we may take an approach of building something very specifically for K-12. Take, for example, the Funster mode of our ePortfolio app. The core portfolio experience is something that is rippling into higher education, and I would imagine is going to be a very big part of how employees demonstrate their skills in the future as well.

So we try to we try to put our energy into building a common learning platform, which does make it a little bit easier for us to serve these different markets as we continue to grow globally.

In our core learning platform, Brightspace, we we try to include everything that you’re probably going to need to deliver an outstanding learning experience. If you compare our core technology to basically the entire suites of most of our competitors, you should actually find everything in there. We’re trying to make it really easy for folks. You just need to worry about one technology, one product, to be able to get up and running and deliver an outstanding experience for students.

Building on Top of the Core Product

John Baker: There are clients that have more demands. So we’ve built in two components that you can actually leverage on top of that platform. One is Engagement Plus. It provides things like gaming engines, the ability for you to do lecture capture, or flipped classroom stuff where you can record a lot of media, different types of interactives, and then be able to engage students differently when you get to a traditional class experience.

Then, on the Performance Plus side, we’ve really built the next generation of artificial intelligence and machine learning to support and identify at-risk students. We’re able to intervene and get them on the right track with some virtual agents. We can also nudge faculty to use their judgment to determine whether the student is on or off the right track, and to provide different remediation or enrichment pathways to help students reach their full potential.

So in that case, we’re applying a lot of the latest AI technologies to improve retention and outcomes for students. It’s having a really positive impact for the clients that are embracing these technologies.

‘Who’ Is D2L?

The above answers some of the ‘what,’ and ‘where’ about the company. But we also asked Baker ‘who’ D2L is.

John Baker: I started D2L in my third year of university. For me, it was answering one key question, and that was: what’s the most important problem that I could solve that would have the biggest impact on the world? I couldn’t think of anything bigger than transforming the way the world learns. So if there’s one overarching mission, that’s it. We want to make sure the learning experiences that folks get at every age are as good as they possibly can be. It doesn’t matter if you’re a kindergarten student in the far north, in a remote community, or if you’re blind, or if you’re deaf, or live with other different learning disabilities, through to you attending one of the best universities in the world. We want to make sure that that learning experience is as good as it can be, so that it improves outcomes.

More recently, we’ve been extending that mission to embrace the future of work. I don’t think any company is going to have all the resources they need in-house to be able to reskill and upskill all their employees as automation and artificial intelligence comes in and displaces jobs. I think our mission has gotten more important. Over the next decade, it will be critical for us to get this right, to be able to support folks as they go from one career to the next, one job to the next, and prepare the next generation of students for that future.

So that’s what we rally around. You you can talk to any D2Ler. They live and breathe that, and have since we started the company.

We also take it down another level and try to actually measure impact. We ask: Are we improving adoption? Are we improving educational outcomes? Are we improving retention? For companies, are we seeing more employees being retained? If it’s a school, are we seeing more students graduate? Are we seeing engagement go up? Satisfaction? Are we seeing productivity come into play?

It’s not just about improving the educational outcomes, we also need to make sure that we make things more productive. Competency-based education is a good example of us living our mission. We’ve been looking for ways to transform the actual experience to have a better impact. And that’s a great example where all the studies that we’re looking at are showing that students are working twice as fast, scoring higher on exams, and retaining the knowledge for longer.

D2L’s Development Process

In this last point, D2L stands out in terms of how they view product development. In a previous interview with eLearning Inside, Instructure CEO Dan Goldsmith discussed the development of AI and predictive capabilities and the march towards personalized learning with the company’s LMS Canvas and Bridge. “In terms of Instructure, we view our organization as a supporting player in the move toward that personalized journey,” Goldsmith said. In other words, for Instructure, the community asks, and the LMS delivers.

According to Baker, however, D2L takes a more proactive approach.

John Baker: I’d say probably 80% of the functionality we deliver is a direct response to us working very closely with our clients to deliver the features or functionality that they were looking for.

There’s another 20% where we’re trying to push the market forward. When we first started, no one was really embracing competency-based education. It was nearly impossible to deliver. But today, many of the cutting edge clients are seeing tremendous growth from that shifted educational model.

As sort of the first wave of learning platforms came into play, it was really around just digitizing the traditional classroom. Things like competency-based learning or leveraging artificial intelligence are a way for us to go beyond that to transform the actual experience to make it better for students and make it better for faculty.

D2L Has Been Building Machine Learning and Predictive Capabilities Into Brightspace for 10 Years

Both the design partnership process and the push into emerging technologies has led D2L to significantly increase the analytic and predictive capabilities of Brightspace.

John Baker: I think when we talk about artificial intelligence, it’s helpful to understand what it actually means. In our case, we tend to use a sub-domain called machine learning. What we’re trying to do is leverage machine learning and artificial intelligence to make the educational experience more human. Machines or AIs are really good at building predictions. But it’s still up to the student or the faculty member to make a judgment call, to take action, to understand what the data actually means.

So, our AIs today are good at predicting things like what grade you’re going to get. We’re very good at predicting which students are at risk, and why, based upon a number of different indicators. So we build things like dashboards for engagement, or we built out what we call our Student Success System, which helps you understand the reasoning behind why a student might be at risk.

So, as a faculty member, you can take an action, you could give them a little nudge, give them a little support, give them some encouragement. You can create different adaptive learning pathways within the system, so that you can have a student pick up the right remediation material. We can identify that foundational building block, if you will, that they need to be successful in the course that they may not have picked up previously.

It’s not about replacing, it’s about providing great tools to help make the right predictions to help folks stay on the right track.

Let’s say someone did poorly on the first couple of assignments in Intro to Calculus. The expectation is they’re probably dropping out, like many others would. And so the nudge could be, ‘Hey, I noticed you got 50% on the first assignment and 60% on the second assignment. Congratulations! That’s progress. 20%, in fact. So here’s what we’re going to be expecting for the third assignment. Here’s some scaffolding that could help you do better. And don’t forget to come talk to me, I’m happy to chat as a professor.’ We could actually automate that based upon a whole bunch of different components within our system.

What we are finding is these things are actually having an impact. Some of our clients are seeing completion rates go from 60% in these large intro courses all the way up to over 90%. If you can give people a reason to persist, then it makes it a lot easier for them to be on track for graduation.

“If you’re not building predictive engines, you’re not providing the right insights or data for folks to make decisions.”

We’ve been working on machine learning and AI for 10 years. We will continue to build technologies to support it, because if you’re not building prediction engines, which is at the core of what these things are doing, then you’re not providing the right insights or data for folks to make decisions.

A sports team would not even think about ignoring analytics or AIs today, but they still have general managers who understand the sport. We’re not replacing the GM. The GM and the coach are still making the calls. But we’re delivering more data at your fingertips, instead of you having to go off and crunch all this yourself as a faculty member or as a student.

How you position it, how you leverage it, how you use it—that’s where the controversy is. Some folks are getting themselves into bit of a mess there. But in our case, it’s all about improving student learning. And if we can do that, then I think we’re in a good spot.

Personalized Learning on the Horizon

Baker says that handing educators better, more informative tools is the main strategy surrounding AI development with Brightspace going forward. But, long-term, D2L has other plans as well.

John Baker: We want to get to a point where we can really truly personalized learning experience, help each person reach their personal level mastery, and help them potentially even go beyond where they even dreamed they could go. So those are all big parts of it. Twenty years from now, I would love to talk about how these things are helping us spark imagination, helping us build better entrepreneurs, helping us build the skills that we’re going to need to be successful in the future of work and the future of living. And AIs will be a big part of how we actually go off into that.

More predictive capabilities means more data being collected about users. And in some of the worst instances, it means incorrect predictions of users that lead to other consequences down the line. These are users who don’t necessarily have a choice as to whether or not they want to use these products.

We asked Baker how D2L weighs data privacy and the user experience against the capabilities it offers to the paying customer.

John Baker: I think people have confused AIs with judgment machines, and these are not meant to be judgment machines, they’re meant to help you make the right call to support improving an outcome. It’s not about them being used to punish someone or to help them achieve a worse outcome.

And as long as we recognize that, as long as we leverage these technologies in a way that we can actually improve education, then we’re going to be in a better spot.

Even in the early days of D2L, faculty could track how long students were online, what progress they had made, and the history of how students are performing across the entire course. There has been all kinds of different data points that have been gathered and used in traditional classrooms or online classrooms. In the early days, we said, ‘Okay, let’s make that same data available to the students.’ So they can see their own progress and they can see how they rank relative to their classmates without being identified.

Back then, 20 years ago, that was a pretty radical idea. So we said we’ll give people the ability to turn that off if they want to. I can tell you: very few people ever want to turn it off.

As we get more comfortable with these technologies helping us, then I think more and more people are going to want to see these things turned on. But if someone wants to be forgotten, then we should give them the right to be forgotten. We’ve always taken that into account at D2L. We have security by design, we try to do privacy by design, we do accessibility by design. These are core things around how you actually build technologies, and they’re going to probably be even more important if folks don’t want to participate in that experience.

But I really do think that’s going to hold them back [if they decide to turn data tracking off]. Because these technologies  are really helping the students get ahead.

Featured Image: RawPixel, Pexels.

Correction September 30, 2019: A previous version of this article reflected John Baker saying that VISA was a corporate client of D2L; that the company had offices in the Netherlands, the U.K., Singapore, Australia, Brazil, and Colombia; and that D2L has clients in nearly 50 different countries. VISA is not a current D2L client. The company employs people in the countries listed above, but does not necessarily have offices in each, and it has clients in nearly 40, not 50, different countries. The article has been changed to reflect this.