There are a lot of concerns in learning and development (L&D) teams right now. These mostly won’t show up in a learning management system (LMS) dashboard, but that doesn’t make them any less urgent. What used to look like a six-week development cycle for a training module on an important new feature is now a three-times-updated relic: a beautiful eLearning module that is mostly outdated by the time it goes live.
The traditional instructional design world was built on a foundation of precision and patience. And to be clear, it was built that way for good reason. Frameworks like ADDIE gave us something invaluable.
They gave us a structured, reliable roadmap for developing learning that is researched, well-sequenced and fit for purpose. We analyzed our learners, designed for their context, developed with care, implemented thoughtfully and evaluated rigorously. This approach wasn’t bureaucracy; it was detail-oriented. And most importantly, it worked.
As we’re seeing more and more changes in learning technologies, we have to consider whether traditional L&D approaches can still keep pace with the avalanche of AI-related changes happening in the industry.
The Velocity Gap
Let’s start by reviewing the scope and importance of the challenges we’re facing. According to Mercer’s 2026 Global Talent Trends report, organizations are already planning how to redesign work around AI and automation. Retraining and upskilling are central to that plan. The World Economic Forum similarly projects that 39% of workers’ existing skill sets will be transformed or become outdated over the 2025-2030 period.
That is an enormous challenge for L&D. And it arrives at exactly the moment when our traditional timelines are most exposed.
When the tools your team uses are updated on a near-weekly basis, a six-to-eight-month development cycle for a “perfect” training course is no longer a luxury; it’s a liability. By the time the course launches, the interface has changed, the workflows have shifted and the best practices have been superseded. You haven’t delivered training. You’ve delivered a historical document.
This is what we might call the velocity gap: the distance between how fast AI is evolving and how fast L&D is currently equipped to respond. Bridging it doesn’t require us to abandon our professional standards. It requires us to evolve how we apply them.
Evolution, Not Replacement
While there is a lot of discussion about change, it is important to note that the foundational principles of good instructional design haven’t actually changed. “Analyze, Design, Develop, Implement and Evaluate” are still the right actions. It is not the process that has changed. It is the speed at which we need to implement it and, thankfully, so have the tools at our disposal.
This is less about replacing ADDIE and more about implementing it at a different tempo, just with better instruments.
A good example of this in practice comes from OpenText. They used an AI-augmented ADDIE workflow to synthesize training materials and learning objectives directly from source documents. By applying those foundational principles through an AI co-pilot, they moved from months of manual development to a system capable of generating full certification exams and interactive modules in much less time.
They did all that without sacrificing evaluation standards. The “Analyze” phase now draws on AI to scan real-time performance data, from support tickets to sales call transcripts, to surface skill gaps in minutes. The “Evaluate” phase has shifted from end-of-course surveys to predictive analytics that track behavioral change the moment a learner returns to their role.
Same principles. Radically different execution speed.
Three Ways to Become an Agile L&D Team
So, what does a more agile approach look like in practice? There are three shifts that matter most.
1. Focus on Principles Over Platforms
The temptation when a new AI tool lands in your organization is to immediately design training around it: its interface, its features, its menu structure. The problem is that all of those things will change. Probably soon.
The more durable approach is to anchor your training in the underlying principles and competencies that allow people to work effectively with any version of the tool. How does this type of AI model generate its outputs? What are the common failure modes? How should a human reviewer approach the output critically? These are the questions that remain relevant even when the interface is updated on a Tuesday with no warning.
This doesn’t mean ignoring the practical, platform-specific content; learners absolutely need it. But if you treat the principles as your foundation and the platform specifics as a layer on top, you’ve built something you can update in a day rather than redevelop from scratch.
2. Prioritize “Good Enough” Over “Perfect”
This is perhaps the most uncomfortable shift for instructional designers who take pride in their craft or their ability to construct highly polished videos. But in an Agile L&D environment, a genuinely helpful microlearning module delivered today is more valuable than a comprehensive, beautifully polished masterpiece delivered eight weeks from now.
The “MVP” mindset applies here. Develop a minimum viable product and build from there. Perhaps start with a short video, a single interactive scenario or a concise job aid. The role of the MVP should be to address the most immediate pain point your learners are facing. Then iterate on it over time as you gather feedback and as the technology continues to evolve. This is not a lowering of standards. It is a recognition that learners need AI skills more urgently than ever and that a useful, imperfect resource beats a perfect one that arrives too late.
One important caveat here: The push for AI skills should not crowd out everything else. AI is a powerful and rapidly growing area of workforce development, but it is not the only thing your learners need. L&D will need to address the most immediate technical needs while continuing to build broader professional competencies. That means other areas, such as communication, critical thinking and judgment, still need attention, as AI is not going to be fully useful in these areas any time soon.
3. Use the Tools at Your Disposal
There is a certain irony in the fact that many L&D professionals are designing AI training without actually using AI in their own workflows. Changing that is also an essential part of adapting to these challenging times.
One simple way to acknowledge this revolution is to build a grounded AI chatbot to handle the repetitive, low-complexity questions that currently swallow your calendar. How do I access the LMS? When do I need to finish the EU AI Act compliance update? These are questions you are already answering, over and over, via Slack messages and emails. An AI assistant with the right grounding documents can handle many of them.
To be clear, this is not about replacing the human element of L&D. It is about reclaiming your time for the work that actually requires your expertise: complex learning design, stakeholder consultation, nuanced performance analysis and creative problem-solving. These are the areas where your professional judgment is irreplaceable. Automating Level 1 questions is what frees you up to do them well.
Beyond the chatbot, the AI toolkit for L&D professionals is growing rapidly. Generative AI can accelerate content drafting, scenario development and even assessment creation. Tools are emerging that can translate a subject matter expert interview into a structured learning outline in minutes. L&D professionals who are experimenting with these tools now will have a significant advantage over those waiting for the “perfect” implementation plan.
The Path Forward
There is a version of this conversation that turns into a referendum on whether ADDIE is still relevant or whether traditional instructional design has a future. That’s not the conversation that should be going on right now. The real question is, how can we ensure our learners have the skills they need at a time when they need them? This question is even more challenging when the idea of a six-month development cycle for L&D is entirely out of the picture.
The best answer to this question is to build an Agile L&D approach. And not just to take that approach as a buzzword, but to integrate Agile as a driving operating principle. It is possible to build modular eLearning courses that can be updated independently. It can also be useful to iterate courses over time rather than pursue perfection and improve only based on real feedback. Overall, the guiding principle should be to provide something useful and on time, rather than something perfect and long overdue.
The goal isn’t to work faster for the sake of being the best and fastest. It’s to close the velocity gap so that when your organization takes its next AI leap, your learners aren’t a training cycle behind. They’re ready.
