Higher Education Has a Data Problem. But It Isn't a Lack of Data.

The next generation of learning technology won’t simply tell institutions what happened. It will help them understand what is happening — and what is likely to happen next.

Higher education does not suffer from a lack of data.

Universities may have more information about their institutions and students today than at any point in their history.

Learning management systems track assignments, grades, and course activity. Student information systems contain enrollment and academic records. CRM platforms track prospective students and communications. Surveys measure satisfaction and sentiment. Advising systems document interactions. Advancement platforms maintain relationships with alumni and donors.

Then there are the thousands of smaller signals generated every day as students attend classes, interact with professors, engage with resources, participate in campus life, communicate with their institution, struggle, recover, succeed, disengage, and eventually graduate.

The problem is that much of this information lives in different places, serves different departments, and is analyzed at different moments.

Universities have become very good at recording what happened.

The bigger opportunity is understanding what is happening while there is still time to change what happens next.

At Artifact Digital, this is a problem we’ve been thinking about in different forms since 2019. Our work in digital experience, research, strategy, learning, and emerging technology kept bringing us back to a deceptively simple question:

What if an institution could recognize the conditions that lead to success or failure before the outcome itself?

That question has become increasingly important as artificial intelligence gives us entirely new ways to think about learning technology. And it has led us toward what we believe could become a new category of higher education technology: the Predictive Learning Platform, or PLP.

The Rearview-Mirror Problem

Consider how institutions traditionally identify a struggling student.

A student begins a semester motivated. Then something changes.

Maybe the material becomes difficult. Maybe the student stops understanding a foundational concept but doesn’t want to raise a hand in front of the class. Maybe work obligations increase. Maybe several missed assignments compound into a larger problem. Maybe the student feels disconnected from the professor. Maybe the course itself has an instructional problem affecting dozens of students simultaneously.

Each event may seem insignificant on its own.

Eventually, however, those small moments accumulate into an outcome. A poor exam grade. A failed course. A withdrawal. A decision not to return the following semester.

The institution eventually sees the result.

The question is: when did the institution actually have enough information to recognize the trajectory?

Those aren’t necessarily the same moment.

By the time a retention report identifies a problem, the most important opportunities for intervention may have happened weeks or months earlier.

That is the rearview-mirror problem.

Traditional analytics are exceptionally useful for understanding what has already occurred. But education happens in motion.

The most valuable question isn’t always “What happened?”

Increasingly, it is: “What appears to be happening right now, and is there something we can do about it?”

From Systems of Record to Systems of Understanding

For decades, enterprise technology has largely been organized around systems of record.

These systems are essential. They give institutions reliable places to store grades, enrollment information, course materials, financial information, communications, and countless other forms of institutional data.

The LMS, for example, was an enormous step forward in organizing digital learning. But an LMS fundamentally answers questions about the mechanics of learning:

What course is the student enrolled in? What assignments exist? What has been submitted? What grade did the student receive?

Those are important questions.

But imagine another layer capable of helping institutions explore questions such as:

Is this student beginning to struggle? Where might understanding be breaking down? Are multiple students encountering difficulty at the same point? Is a course trending toward an unusually high failure or withdrawal rate? Is engagement improving or deteriorating? What patterns tend to precede successful outcomes? Where could an advisor, professor, TA, or institution make a meaningful difference?

Those are different kinds of questions.

They require us to move beyond systems that simply record learning toward systems capable of helping institutions understand learning as it unfolds.

That distinction is at the heart of how we think about a Predictive Learning Platform.

A University Is Not a Collection of Departments

One of the biggest conceptual problems with university technology is that technology often mirrors the organizational chart.

Admissions has its systems. Faculty has theirs. Student services has another. Advancement has another. Marketing has another. Institutional research has another.

Students experience something completely different.

To a student, there aren’t twelve departments and fourteen software platforms. There is one university.

A student’s experience is continuous. The prospective student researching a university eventually becomes an applicant. The applicant becomes an enrolled student. The student enters classrooms, builds relationships, encounters challenges, participates in campus life, receives advising, communicates with professors, and hopefully progresses toward graduation.

That graduate becomes an alumnus. That alumnus may become a mentor, donor, employer, advocate, parent, or community leader.

Families experience another version of that journey. Faculty experience another. Administrators see another.

These aren’t separate stories. They are interconnected perspectives on the same living institution.

Yet our technology frequently separates them.

What if instead we thought about the university as a living ecosystem of relationships, experiences, behaviors, and outcomes?

That shift in perspective opens up something much more interesting.

The University as a Living Feedback Loop

A metaphor that has influenced our thinking is a loom.

A university contains thousands — sometimes hundreds of thousands — of individual threads. Students. Professors. Teaching assistants. Advisors. Administrators. Families. Alumni. Courses. Programs. Resources. Interactions. Experiences.

Every thread moves differently. But together they create the fabric of the institution.

The challenge is that universities often examine individual threads without being able to see the larger pattern forming in real time.

A Predictive Learning Platform changes the question. Instead of simply collecting more information, the goal becomes connecting meaningful signals into an active feedback loop.

See. Learn. Respond. Adapt. Then repeat.

The institution becomes capable of learning about itself while learning is happening.

That has profound implications.

Imagine Knowing Earlier

Imagine a professor beginning a semester with 120 students.

Traditional systems might eventually show that 18 students failed the course. That’s useful information.

But imagine knowing during week three that a particular concept appears to be creating confusion across a significant portion of the class.

Now the professor has options. Clarify the concept. Introduce another example. Provide additional material. Ask a TA to host a focused session. Change how the next lesson is structured.

The goal isn’t to replace the professor’s judgment. It’s to give the professor better visibility.

Now imagine the same capability at the student level.

A student doesn’t suddenly become “at risk.” Risk develops. Engagement changes. Understanding changes. Behavior changes. Patterns emerge.

If those patterns can be recognized earlier, the institution has an opportunity to respond while the student’s story is still being written.

That is fundamentally different from generating a report saying what percentage of students failed.

Prediction creates the possibility of intervention. And intervention creates the possibility of a different outcome.

This Isn’t About Predicting Failure

It’s easy to hear the word “predictive” and assume the objective is to create increasingly sophisticated mechanisms for labeling students.

That would be a terrible use of this technology.

The purpose of prediction should not be to decide someone’s future. It should be to create more opportunities to change it.

A prediction should never become a verdict.

A student identified as potentially struggling isn’t a failing student. They’re a student for whom the institution may have an opportunity to help.

A course trending toward poor outcomes isn’t necessarily a bad course. It may be revealing a particular instructional challenge that can be addressed.

This distinction matters enormously.

AI should not remove humanity from education. Used thoughtfully, it should give humans better opportunities to show up for one another.

Professors can teach. Advisors can advise. Administrators can allocate resources. Students can understand where they stand. And institutions can respond sooner.

Technology becomes valuable not because it automates the relationship, but because it can help strengthen it.

The Missing Feedback Loop

There is another important piece of this.

Learning isn’t simply something institutions deliver to students. Students continuously generate information about how well the institution is teaching.

Every moment of confusion contains information. Every breakthrough contains information. Every disengagement contains information. Every question contains information. Every successful course contains information. Every struggling course contains information.

Historically, much of that information disappears.

Sometimes institutions capture pieces of it through end-of-semester evaluations or annual surveys. But imagine running a business where you could only hear from your customers once or twice a year. We would consider that an enormous strategic disadvantage.

Education should increasingly move toward more continuous forms of understanding.

Not constant surveillance. Not endless surveys. Not more administrative burden.

Better feedback.

The difference is critical.

The AI Opportunity Is Bigger Than Chatbots

Much of the current conversation about AI in education has focused on generative AI.

Can students use ChatGPT? How should professors handle AI-generated assignments? Should universities build AI tutors? How can faculty use AI to create course materials?

Those are important conversations. But they represent only one dimension of the transformation underway.

The larger opportunity may be institutional intelligence.

What happens when AI can help an institution recognize meaningful patterns across enormous amounts of information that no individual administrator, professor, or advisor could reasonably process? What happens when those insights become available at the moment they are useful rather than six months later? What happens when the institution itself becomes capable of continuously learning?

That’s a much larger idea than adding a chatbot to an LMS.

Why We’ve Been Thinking About This Since 2019

Artifact Digital didn’t arrive at this idea because generative AI suddenly became popular.

We’ve been thinking about the relationship between digital experiences, human behavior, data, and outcomes since 2019. Our work has consistently focused on understanding people within complex systems.

Where do they struggle? What are they trying to accomplish? What signals indicate friction? What information is missing? What patterns emerge across an experience? How can technology respond more intelligently?

Those questions apply to commerce, enterprise software, digital ecosystems, and increasingly to education.

AI has dramatically expanded what is possible, but the underlying challenge hasn’t changed: technology is most valuable when it helps us understand people well enough to create better outcomes.

Our work around the Predictive Learning Platform is an extension of that philosophy. We aren’t interested in putting AI into education simply because AI is available. We’re interested in what becomes possible when human-centered experience design, institutional knowledge, behavioral signals, and artificial intelligence begin working together responsibly.

Exactly how those pieces come together is something we’re continuing to develop at Artifact. But the destination is becoming increasingly clear.

From Student Success to Institutional Success

The implications extend beyond individual students.

A university is itself a learning organism. Courses change. Programs evolve. Student populations change. Economic conditions change. Technology changes. Expectations change.

Institutions need to understand not only whether students are succeeding, but why.

What creates persistence? What creates belonging? What creates academic momentum? Which interventions actually work? Where does institutional friction appear? Which experiences strengthen the relationship between a student and the university? Why do graduates remain connected to one institution for decades while others disappear immediately after commencement?

Those questions stretch from enrollment through graduation and into alumni relationships.

A university that can continuously learn from those experiences becomes better equipped to adapt. And adaptability may become one of the defining characteristics of successful institutions over the next decade.

Toward the Predictive University

I believe we’re approaching an important transition.

The first generation of university digital transformation was largely about putting information online. The next connected systems and digitized workflows. Then came analytics, dashboards, and increasingly sophisticated reporting.

Now AI gives us the opportunity to take another step.

From recording to understanding. From reporting to recognizing. From reacting to anticipating. From isolated systems to connected intelligence. From periodic measurement to continuous feedback.

The Predictive Learning Platform is our exploration of what that future could look like.

Not a replacement for professors. Not another dashboard demanding attention. Not an algorithm deciding which students will succeed.

Something much more human: a way for institutions to listen better. To recognize patterns earlier. To understand where people are struggling. To see where they’re thriving. To strengthen the relationships connecting students, faculty, families, administrators, and alumni.

And ultimately, to give institutions the ability to act while action can still make a difference.

Because the most important insight about the future of a student isn’t knowing what will happen. It’s knowing what might happen early enough to help change it.

That is the future we’re interested in building.