Here is a design problem I keep turning over: how do you take a two-day in-person training and make it work asynchronously, across five sites, each with its own context, pace, leadership, and starting point? The temptation is to treat it as a content problem. Record the sessions, build the slides, add a quiz, and call it a course. But a two-day training rarely works because of the content alone. It works because of what happens in the room, and because of the follow-up that usually comes after. Strip those away and you are left hoping that five very different sites all translate the same material into practice on their own. They won't, and it isn't a motivation problem when they don't. So the first move is to build the developmental coaching in from the beginning, not as an add-on once the "real" training is done. From the start, participants know what application is supposed to look like, what their coach will be looking for, and how the follow-up is meant to help, not to catch them out, but to identify where progress is happening and where more support is needed. The coaching is part of the design, not a bolt-on at the end. Then, a later check-in refreshes the learning and looks at how it is actually showing up in everyday practice. Participants won't necessarily know the exact date someone will observe, but they are not left guessing about the purpose or the framework. The why and the how are transparent. The point is not to surprise anyone. It is to see the work in its real conditions. And in that check-in, the coach is looking at a few things at once: what is working, where application is difficult, and what might explain that difficulty. Does the person need more practice? Clearer guidance? Resources they don't have? Or is it the conditions around them that need to change? That diagnosis is what guides the next steps the two of them work on together. The answer is rarely "try harder." It is usually more specific and more useful than that. What I like about designing it this way is that the five different contexts stop being a problem to flatten and become information the system can actually use. Bring me the specific implementation problem your school or team is facing: https://www.themultilingualmind.com/services