Every learning platform seems to be racing to add AI features, from chatbots to content generators to personalized dashboards. But are all of these tools actually helping learners, or are they just AI for AI’s sake?
We believe that the same principles that shape excellence in digital learning apply to AI in learning: High Engagement at Scale™. Using the principles of High Engagement at Scale™ to lay a strong strategic foundation helps avoid the risk of building flashy technology that looks innovative but fails to deliver meaningful impact for learners and organizations. It’s often cited that 95% of AI pilot initiatives get zero return. Here’s how we apply High Engagement at Scale™ to reach that successful 5% when creating AI features for learning.
How High Engagement at Scale™ Sets the Stage for Successful AI Applications
When organizations rush to implement AI in learning, they often start with the technology and work backward. They ask, “what can AI do?” rather than, “what do our learners need?” This approach leads to solutions searching for problems. It’s the classic example of what not to do in product design.
We developed High Engagement at Scale™ from our experience building over 350 learning experiences over the course of 15+ years. Grounding AI decisions in evidence-based principles that have been pressure-tested across hundreds of learning experiences ensures that AI investments deliver simultaneous learner engagement and program scale. Scale is where AI really shines, enabling previously impractical pedagogical approaches. Let’s examine examples of using High Engagement at Scale™ to guide decisions on what AI tools to build.
AI for Personalized Learning Pathways: Learner-Centered Content
The first component of our High Engagement at Scale™ framework is Learner-Centered Content: high-quality, relevant content that’s scaffolded and chunked appropriately. We spend a lot of time with our clients on content strategy, mapping course content to learning personas, and determining the right content modalities. We’ve been creating personalized learning pathways for years, but using AI takes personalization to the next level. With AI we can go beyond pathways geared towards particular learner personas to pathways created for an individual learner.
One use case is using AI for learning experience navigation. For example, let’s say I need a refresher on business applications of regression analyses. The AI search can direct me to the exact module I’m looking for and tailor the outcome to my professional level of learning. No more wading through a basic statistics curriculum or scrolling through course videos to find the one piece of information I need. Here AI provides just the right content, at just the right time, tailored to a learner’s profile.
In a similar vein, we prototyped an AI chatbot that engages a learner in conversation, then goes beyond just recommending courses or modules. It creates a custom curriculum based on the learner’s needs. So if you say, “I’m a small business owner who needs to learn about finance” you’ll get a collection of short-form videos pulled from across the course catalog, all geared to small business owners.
Using AI for personalized learning pathways avoids one of the major risks of AI: hallucination. The content itself is not AI-generated. Instead, we use AI to navigate vetted, high-quality, human-created content.
AI for Immersive Practice: Active Learning
We all know the saying about practice making perfect. Using AI-powered simulations enables learners to build hard-to-practice skills with realistic scenarios in a safe, low-risk environment.
One example of this is an AI chatbot we helped a client design to teach negotiation skills. Learners practiced negotiating with an AI partner that responded dynamically to their choices, creating a realistic, immersive experience. Importantly, this AI negotiation tool was used to help learners rehearse difficult, high-stakes interactions in a low-risk environment to prepare for a simulated negotiation with real people. The AI tool helped make the person-to-person interaction more valuable. This exemplifies a balanced approach to using both AI simulations and person-to-person interactions in learning experiences.
We’ve also applied this approach to AI chatbots to help financial advisors practice challenging client conversations. The same principles apply: the ultimate interaction is in-person. The AI tool provides a new, immersive modality for advisors to experiment, rehearse, build confidence, and learn. It does not replace human interaction.
Productive Failure in Safe Spaces: Unbounded Inclusion
While it’s easy to understand the value of practice (the Active Learning component), the “safe” aspect is an often overlooked critical component that ties into Unbounded Inclusion. One of the most powerful learning mechanisms is productive failure: struggling with challenges, making mistakes, receiving feedback, and trying again. However, traditional learning environments often make failure feel risky. With careful construction, AI simulations provide a judgement-free space for all learners. This type of inclusion enables deep learning because it makes it “safe” to fail. It’s a key component to the success of AI tools like the negotiation or financial advisor training chatbots.
Socratic AI Tutors for Active Learning
One important application that’s gaining traction is the AI Socratic tutor [link]. An AI Socratic tutor guides learners through problems by asking questions rather than providing direct answers. This tool avoids the risk of cognitive offloading, a real risk when deploying AI in learning. OpenAI’s “Study Mode” is perhaps the best known example of this approach. We first wrote about Socratic tutors in early 2024, and it’s been encouraging to see their adoption and growth among AI leaders.
When a learner struggles with a concept, an AI Socratic tutor doesn’t simply jump to the answer. Instead, it asks probing questions that help the learner identify gaps in their understanding, make connections to prior knowledge, or consider the problem from different angles. It provides hints and scaffolding calibrated to each learner’s needs, gradually reducing support as competence grows. Socratic tutors demand reflection, a key component of Active Learning.
Socratic AI tutors allow every learner to have access to coaching support whenever they need it, scaling the benefits of one-on-one guidance that would otherwise be a huge burden to the instructor. And this is one of the core benefits of AI in learning: scaling previously impractical approaches like one-on-one tutors for all learners.
Strategy Over Hype: Scaling Without Sacrificing Engagement
Technology alone doesn’t create world-class learning experiences. That requires a firm understanding of learning design that’s then thoughtfully executed with top-notch technology.
From this perspective, AI is just another technology with which to execute a learning strategy, albeit an enormously powerful one. AI’s power can make sophisticated pedagogical approaches accessible to millions of learners. But, it also has the power to outsource thinking and, therefore, meaningful learning. And it can also just flop–become yet another app that nobody uses.
By holding to proven practices like those of High Engagement at Scale™, you can avoid the risks of detrimental or merely unappealing AI tools and create AI that powers meaningful learning at a previously unimaginable scale.
