AI in Workplace Learning Has a Content Problem: Just Not What You Think

TalentLMS’ Thanos Papangelis offers this commentary on how AI in workplace learning has a content problem. This article originally appeared in Insight Jam, an enterprise IT community that enables human conversation on AI.

For the past two years, the conversation around AI in workplace learning has centered on content. How quickly can we create it? How much can we personalize it? How much time can we save? Those were the right questions to ask at the start. But they’re becoming the wrong questions to focus on now.

The first thing AI did for workplace learning was make content easier to produce. Learning teams were buried building courses, updating materials, and keeping everything current as products and policies changed underneath them. AI cuts that work down dramatically. Programs that once took weeks to build can now be turned around in days.

So naturally, the industry called it a revolution and moved on.

Except the L&D leaders I talk to haven’t moved on. They’ll tell you the content creation problem is largely handled and then almost immediately start talking about something else. Something harder. Producing training faster doesn’t actually mean people are learning more.

That gap has been sitting there the whole time. AI didn’t create it. If anything, AI has made it harder to ignore. When creating content becomes easy, the bottleneck shifts. The challenge is no longer production; it’s whether learning translates into workforce capability and measurable skills that actually improve performance.

Capability, Not Just Completion

Here’s what it looks like in practice. An employee completes a compliance module. The LMS records 100% completion. Three weeks later, put them in a situation that requires applying what they supposedly learned, and there’s a real chance they’ll hesitate, guess, or get it wrong. Not because they weren’t paying attention, but because watching information go by and actually processing it are completely different cognitive activities.

Learning scientists call what’s missing active processing: asking questions, testing assumptions, making low-stakes mistakes, and getting immediate feedback. This is how people build durable capability. Not by consuming content, but by engaging with it. Research on retention has pointed to this for decades. Without opportunities to practice and apply, people forget most of what they’re taught within days.

Organizations don’t have a completion problem as much as they have a skills visibility gap. Completion data tells you who finished training. It doesn’t tell you whether employees can perform. Too often, critical capabilities remain hidden, leaving leaders with limited skills visibility and little understanding of where their workforce is thriving or where important skills gaps still exist.

Enterprise training has almost never delivered active processing at scale. We’ve compensated with managers who coach, mentors who answer questions, and instructors who facilitate discussions. Those relationships matter enormously. But they’re constrained by time, geography, seniority, and who happens to be in the room. Not everyone gets the same access. As organizations grow, those differences become even more pronounced.

AI Is Moving Out of the Back Office

For most of its early life in L&D, AI was a production tool. It helped teams create and manage content. It didn’t interact directly with learners. That’s changing.

The more interesting application, and the one with the greatest long-term potential, is AI-powered skill development embedded directly into the learning experience. Instead of simply delivering information, AI can help employees actively engage with it while they’re building capability.

A manager preparing for a difficult performance conversation can work through it beforehand. Not by reading tips about giving feedback, but by practicing, receiving responses, and refining their approach. A customer-facing employee heading into a challenging renewal can rehearse different scenarios before the meeting. Someone onboarding into a technical role who encounters an unfamiliar concept doesn’t have to wait until next week’s office hours. They can ask questions, explore the topic, and deepen their understanding at the moment they need it.

Those interactions create opportunities for self-led skill building. Employees aren’t waiting for the next formal training event or relying entirely on a manager’s availability. They can practice, receive feedback, and continue improving through AI-supported progression that reinforces learning over time.

None of that replaces real experience or human judgment. I want to be clear about that because the conversation too often swings between extremes: either AI is going to replace coaches and managers, or it’s useless without them. Neither view reflects what’s actually happening. The real opportunity exists in the space between formal learning events. Today, most employees complete training and return to work with little support connecting what they learned to the situations where they’ll actually need it.

That’s the gap. It’s not a technology problem. It’s a practice problem. AI simply gives organizations a scalable way to help fill it.

What Buyers Should Be Asking About Now

For organizations evaluating learning platforms, this shift changes which questions matter most.

The first wave of AI in learning management systems focused almost entirely on efficiency. How quickly can we build content? How much administrative work can we eliminate? Those remain valuable questions, but they’re no longer enough.

Organizations should start shifting from completion metrics to capability metrics. The more important question is whether learning is changing how someone performs weeks after training ends. Can organizations build capability they can actually measure? Can they understand the real skills impact of their learning investments instead of simply tracking course completions?

There’s also an important trust dimension. Using AI to draft learning content that humans review before publication carries one level of risk. Using AI to interact directly with employees throughout their learning journey is something else entirely. The quality of those interactions matters immediately. How is the AI guided? What guardrails are in place? How does it support greater skills clarity instead of simply generating more conversations? Those aren’t theoretical governance questions. They’re practical questions every organization should answer before deploying learner-facing AI at scale.

The Longer Arc

I’ve been in learning technology long enough to remember when the industry’s biggest promise was access. Put the right information in front of the right people at the right time, and learning would naturally follow. An entire generation of learning platforms was built around that idea: LMS platforms, content libraries, video courses, and microlearning. It solved a real problem. Sharing knowledge at scale used to be difficult. Today, it isn’t.

But somewhere along the way, access became the goal instead of the starting point. That’s how organizations end up with enormous content libraries, impressive completion rates, and very little understanding of whether employees have developed the capabilities the business actually needs. Without meaningful skills visibility, hidden skills remain undiscovered while critical skills gaps continue to grow unnoticed.

What learner-facing AI introduces, if it’s implemented thoughtfully, is a different kind of value. Not faster content or simply more content, but AI-powered skill development that helps shorten speed-to-skill and strengthen workforce capability over time.

This shift is one reason we’re seeing growing interest in learner-facing AI experiences, including TalentLMS’s Learning Playground, which was designed to let employees ask questions, test ideas, and work through scenarios rather than passively consume information. The goal isn’t to replace courses. It’s to make the space between them more productive by giving learners more opportunities to practice, receive feedback, and build capability in context.

That’s not a small improvement. It addresses one of the biggest challenges workplace learning has faced for years: helping organizations move beyond measuring participation and toward understanding whether learning is actually producing the skills that drive business performance. AI finally gives us a scalable way to make capability visible, not just completion.