Mindsmith has been building the agent and medium for learning for years now. In that time, we’ve talked about adaptive learning countless times around the office, everything from casual hallway conversations to genuinely heated debates about whether we should build it at all, and whether the timing was right. For years, those debates ended the same way: not yet. Adaptive learning gets treated like the holy grail of learning. Here’s the story of why we waited, and why now is the time.
Every time we sat down to build adaptive learning, we realized our time was better spent improving the learning content itself. So long as people were still learning from PDFs and article-style content, there was far more progress to be made by giving learning designers better tools to build great linear exercises than by adapting mediocre ones. We also worried that poorly implemented adaptive learning could make things worse. If we didn’t have a learning experience we were 100% proud of, adapting it would just add friction. Adaptation can also spread designers thin across countless hypothetical scenarios.
Isn’t the best way to learn with content adapted to you? The most common version of that claim is that you should teach to each learner’s preferred “learning style.” But it doesn’t hold up. The idea that people learn better when content is matched to their learning style is widely considered a myth. And the problem runs deeper than that one theory: adapting to the individual isn’t automatically better than teaching everyone well. We shouldn’t completely disregard learner preferences, but you’re almost always better off taking a holistic approach. Take a visual learner. You could lean all the way into their supposed preference and serve them nothing but videos and diagrams, but if those videos are formulaic, they’ll just doze off and learn nothing. Even for someone who “prefers video,” you’d be much better off giving them a blended mix of refined exercises.
We tend to box ourselves into thinking eLearning means page after page of text, images, and interactive elements. Almost every instructional design team has been through a love-hate relationship with that format, swinging hard toward interactive ambitions and then retreating to the safety of a video and a quiz when the experience didn’t hold up. eLearning is certainly all of those things, but it can be so much more. Mindsmith has come to define eLearning as anything a learner can do on their own behind a computer. Where classic eLearning mostly asks the learner to sit and consume, we’re now focused is on having them actually do the thing they’re trying to learn, imagine: a short video to introduce and motivate the topic, a matching activity to get their mind moving, a table that lays out how the concept breaks down, then real practice through an AI conversation or a scenario experience, all rounded out with a quiz to make sure it landed. Recently, we’ve focused on giving designers the tools to finally incorporate competency-building into their lessons, with major investments in AI-driven conversational learning and simulation experiences. Mindsmith’s engineering team is four times the size it was a year ago, and we can finally solve these things and make your learners love eLearning.
If you sit down and imagine the ideal learning experience with no limitations at all, you’d probably picture an individualized tutor. But a tutor on their own gets blown around by whatever the learner happens to feel like that day. What makes administered learning powerful is exactly what that free-floating tutor lacks: it stays consistent, it keeps the learner building, and it’s anchored by a rock-solid curriculum and clear objectives. So the real ideal is a tutor with that structure behind them, able to call on any learning experience the learner needs, backed by a team of expert instructional designers. This ideal wouldn’t be just a tutor talking, either. The tutor would show you things, like an animated video that makes a hard topic click, and most importantly, it would have the learner learn by doing the very thing they’re trying to learn. You might take little detours in a session, but the learning program keeps you on course. We imagine that instructional designers become the architects of that underlying structure.
As eLearning becomes able to teach more and more valuable skills even more efficiently, learning teams become more important to the organizations they serve. We’ve seen this pattern before. Despite everything people predicted about AI coding tools, demand for software engineers is actually up, with the work shifting toward higher-value judgment. With Mindsmith, the effect could be even more profound, because designers aren’t just getting a “sous-chef” assistant in the agent, they’re getting a whole new medium to express learning in, the kind of shift the iPhone set off for software.
All of that said, Mindsmith is going all in on adaptive learning, on the smallest end Mindsmith is live worldwide with adaptive conversations. Today we’re releasing lesson branching in closed beta, and we have a series of releases over the coming months that build out the full Mindsmith mission: to elevate the world’s learning.