The Course Was Ready Before The Learners Were
A customer service team needs training on a new returns policy before the end of the week. On Monday morning, the L&D manager uploads the policy into an Artificial Intelligence (AI) tool. By lunch, she has a course outline, a video script, customer scenarios, a quiz, and a facilitator guide.
The course launches two days later. Completion reaches 96%, and most employees pass the quiz on the first attempt. Then the difficult cases start arriving. Agents hesitate whenever the policy falls outside the examples they practiced, supervisors answer the same questions repeatedly, and quality reviews reveal the same mistakes.
The course had been produced faster than ever. The learning had not.
AI Speeds Up Course Creation. It Does Not Speed Up Working Memory.
Artificial Intelligence has transformed one part of workplace learning with remarkable speed. Tasks that once took Instructional Designers days or weeks—drafting outlines, writing scripts, creating scenarios, or adapting content—can now happen in a single afternoon. That shift removes one bottleneck: producing learning content.
Producing a course means organizing information. Learning means building understanding, connecting new ideas to existing knowledge, and creating memories that can later be retrieved and applied. Speeding up one process does not automatically accelerate the other.
Learning science helps explain why. John Sweller's cognitive load theory begins with a simple constraint: working memory can process only a limited amount of new information at one time. Learning depends on building mental structures over time—not on presenting more information more quickly. For L&D teams, the design challenge has shifted:
- Before AI
How do we create enough content? - After AI
How do we decide what deserves a learner's limited attention?
AI has dramatically expanded what can be produced. Human working memory still determines what can be meaningfully processed. If processing information were enough, faster course creation would naturally produce better learning. Research suggests something different.
Learning Still Needs Retrieval, Not Just Exposure
Understanding a concept during training is only part of the learning process. The harder question is whether that understanding survives long enough to influence decisions days or weeks later. Durable learning depends on two conditions:
- Learners actively retrieve information rather than simply review it.
- They encounter that information again after time has passed.
Research helps explain why. Psychologists Henry Roediger III and Jeffrey Karpicke found that actively retrieving information strengthens long-term retention more effectively than simply reviewing the same material again. Testing, in this context, serves a second purpose. It is not only a way to measure what learners know but also a way to reinforce what they are likely to remember.
A large review by Cepeda and colleagues reached a complementary conclusion. Learning becomes more durable when practice is distributed across time rather than concentrated into a single session. Revisiting material after some forgetting has occurred strengthens retention more effectively than covering everything at once.
This creates an interesting contrast with the strengths of Artificial Intelligence. The most obvious use of AI in L&D is speed: generating more content in less time. Human memory works differently. It benefits from interruption, retrieval, and repeated encounters with the same ideas over time.
For L&D teams, that changes the role AI should play in learning design. The goal is to create a sequence of learning experiences that brings important knowledge back at the moments when memory begins to fade.
The Real Improvement Started After Launch
A software company needed to roll out a new customer data policy before updated compliance requirements took effect. Using AI, the L&D team built the training in less than two days instead of the three weeks the process had previously required. The launch looked successful. Nearly every employee completed the course, and quiz scores averaged above 90%.
A month later, quality reviews revealed a different problem. Support agents still struggled whenever customer situations fell outside the examples used during training. The issue was not initial understanding. The issue was retrieving and applying the policy when real work became more complex. Instead of adding more content, the team changed the learning cycle:
- Instead of building a complte course, they added spaced follow-up activities.
- Instead of asking learners to review, they asked learners to retrieve information.
- Instead of ending learning at completion, they extended learning through coaching and practice.
The course itself changed very little. AI was used to generate short retrieval activities—a scenario three days later, another a week later, and a role-specific decision exercise during coaching sessions. The biggest improvement came from making repeated retrieval practical without creating substantially more work for Instructional Designers.
The Course Is Only The Beginning
As content generation becomes faster and easier, the competitive advantage is shifting away from producing information and toward designing learning experiences that help people remember and apply it. Use AI across the learning cycle, not only at the start:
- Before training
Identify what deserves learners' attention and remove information that does not support performance. - During training
Generate realistic scenarios, examples, and decision points that require learners to think, retrieve, and apply new knowledge instead of simply reviewing it. - After training
Create retrieval prompts, spaced practice, coaching questions, and role-specific application checks that bring important knowledge back after formal training ends.
When AI supports the entire learning cycle, its value extends well beyond producing courses. It helps Instructional Designers create repeated opportunities for learners to retrieve, apply, and strengthen knowledge over time.
Artificial Intelligence is exceptionally good at reducing repetitive design work. That creates an opportunity to invest more time in the parts of learning that still require human judgment: deciding what matters most, when learners should practice, and how new knowledge connects to everyday performance.
Conclusion
Working memory remains limited, durable learning still depends on retrieval and spaced practice, and performance continues to develop long after a course is complete. The real opportunity for L&D teams is using the time AI saves to design the learning experiences that happen before, between, and after formal training. When AI supports the entire learning cycle—not just content creation—it becomes more than a productivity tool. It becomes a way to make learning itself more effective. AI can build a course in minutes. Learning still takes time.
References:
- Cepeda, N. J., H. Pashler, E. Vul, J. T. Wixted, and D. Rohrer. 2006. "Distributed Practice in Verbal Recall Tasks: A Review and Quantitative Synthesis." Psychological Bulletin 132 (3): 354–80. https://doi.org/10.1037/0033-2909.132.3.354
- Roediger, H. L., III, and J. D. Karpicke. 2006. "Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention." Psychological Science 17 (3): 249–55. https://doi.org/10.1111/j.1467-9280.2006.01693.x
- Sweller, J. 1988. "Cognitive Load During Problem Solving: Effects on Learning." Cognitive Science 12( 2): 257–85. https://doi.org/10.1207/s15516709cog1202_4
