Earlier this month, we ran a roundtable with L&D practitioners, consultants, and product managers from across Europe. The question we put to the room: if we use AI to do the same things we've always done, but faster, what opportunities do we miss?
The answers were more honest than we expected.
Most organizations using AI in L&D right now are doing the same thing: taking their existing storyboard-to-SCORM workflow and speeding it up. More content in less time. Same format, less effort.
Mark, an independent learning technology evaluator who has worked with a broad range of L&D tools, put the problem directly during the session:
"We can produce more, faster. But it doesn't actually necessarily link to the realities of what people are trying to achieve. What performance impact we're trying to have."
The issue is structural. More content means more to navigate. If learners couldn't find what they needed in a course library before, adding more courses doesn't fix that. It makes it worse.
Wil, a digital transformation and learning consultant who co-facilitated the roundtable, framed it this way: "A lot of organizations use AI to do the same thing we've always done. The question is what gets missed."
His answer: the richness of subject matter expertise. By the time knowledge passes through a traditional L&D workflow and becomes a storyboard or a slide deck, most of the texture is gone. The edge cases, the examples, the context a learner might actually need.
Before you can fix how you render learning content, you have to fix where you keep your knowledge.
Kanika, who works in learning and development for a publishing company with nearly 200 years of history, described the reality in her organization:
"The data doesn't sit in a structured format in any one place. It exists in people's brains, in legacy systems, in offline documents."
Mark added something most L&D teams know but rarely say out loud:
"We can pull it all from Confluence. Yeah, but you don't know what is valid in Confluence and what is not. We've got all these presentations on Google Drive. But 10 of them are a copy of the first one that's been tweaked slightly. Which one is the single source of truth? Curation is a huge challenge."
This is where a common shortcut fails. Thomas, founder of Netresearch and the team behind ScormIQ, described a pattern he's seen repeatedly:
"The typical IT approach is: we just connect Jira, Confluence, the wiki — and then everything will be fine. But even in organizations with a very well-curated document management system, the missing piece isn't the information. It's the knowledge layer."
The knowledge layer is what helps someone navigate and use the information they already have. It's the context, the structure, the understanding of which document to read first, and why. Without it, even a well-organized knowledge base is a locked room.
Wil outlined a three-stage framework during the session that reframes how L&D teams should think about their work.
Capture — Record subject matter expertise richly, before it gets filtered into slides. Interviews, recordings, observations of people doing their jobs. The goal is to keep the texture, not just the conclusions.
Store — Put it in a format machines can read. Not necessarily a full relational database. Wil described his own approach: JSON files on a shared drive, compliant with existing systems, no heavyweight infrastructure needed.
"That really unblocked a lot of progress."
Render — Let the AI produce whatever format the learner actually needs at that moment. A short answer. A translated explanation. A chatbot response on a phone. Or yes, a SCORM module, if that's what the situation calls for.
The shift is subtle but important. The SCORM course stops being the endpoint. It becomes one possible output among many.
Simon, a product manager who has been building internal AI learning tools at his company, described where this approach has already worked in practice:
"Rather than static knowledge you must onboard yourself, participants leave with a working piece of software that helps them in their work immediately. That feels different."
Even organizations that have done the hard work of collecting and storing knowledge often hit a wall.
Thomas described a customer with a sophisticated, well-maintained document management system:
"The information was stored and classified. But for a newcomer, it was not easy to make use of it. There was a very dense silo of information, but the knowledge layer was missing."
AI tutors that understand this distinction can help. Not by replacing human knowledge, but by making it reachable at the right moment. At Sika, for example, we connected ScormIQ to their existing SCORM courses covering complex technical content. Sales engineers can now ask questions mid-course, in their own language, and get answers grounded in the course material. The knowledge was always there. The access wasn't.
One of the more practically useful parts of the session was when the group pushed back on perfectionism.
Thomas's advice was direct:
"My recommendation is to start small. Build a simple idea. You have an inbox where you put all your raw data. Then maybe in a later stage, a little agent that always goes into the inbox, reads everything, and suggests where to put it. So over time it builds your knowledge base. It will not work 100% in the first try. But you will see that you get quite good results."
Ceri, a freelance learning specialist working on an AI integration project for a large technology company, named the real obstacle:
"The letting go of how things are done is really difficult. It's not that people aren't interested. It's the ownership, the architecture, the supporting infrastructure."
Ten percent better is still progress. You don't need a complete system before you start.
The faster-horse trap isn't a technology problem. It's a framing problem. The question isn't "how do we use AI to produce more content?" It's "how do we use AI to make the knowledge we already have more useful?"
That starts with thinking in three stages: Capture richly, store accessibly, render at the point of need.
If you want to see how ScormIQ fits into this kind of workflow — starting with your existing SCORM courses — we're happy to walk through it with you. Book a demo.
---
This post draws on a roundtable discussion held on 14 April 2026 with L&D practitioners from across Europe. Participants included Wil, Mark, Kanika, Simon, Ceri, and others.