Industry trend reports this year keep saying the same thing. AI-driven coaching is going mainstream. Wearables track your heart rate, your sleep, your recovery scores, and feed all of it into an algorithm that adjusts your training plan in real time.
I don't think that's a bad development. Data that used to take a coach years of experience to estimate by feel is now sitting on your wrist, and using it well can genuinely sharpen how a program gets built. I'd rather work with more information than less.
But I want to be honest about where the limit is, because I think it gets glossed over in most of the coverage I've read.
An algorithm can tell you your heart rate variability dropped last night. It can't tell you that it dropped because you had an argument with your sister and haven't processed it yet, and that no amount of programming adjustment fixes that, only naming it does. It can tell you your squat depth changed by a few degrees. It can't tell you that the change showed up the same week you started a new job and your nervous system has been quietly bracing for a threat that isn't physical.
A number is a symptom, not a story. My job, most sessions, is finding the story underneath the number. That takes watching someone move, asking questions that have nothing to do with reps or sets, and noticing things a sensor isn't built to notice, like the half second of hesitation before someone steps under a bar, or the way their breathing changes when a certain exercise comes up.
There's also a harder truth in here. Data can flatten a person into a trend line, and trend lines invite you to chase them. If your recovery score is good, it's tempting to push, whether or not that's actually what your day calls for. Good coaching sometimes means overriding what the data says, because a number from this morning doesn't know what's coming this afternoon, and it definitely doesn't know what you're carrying that never shows up in a heart rate reading.
None of this means skip the wearable or ignore the algorithm. Use the data. I use it too, when clients bring it in. But treat it as one more thing to notice, not the final word on how someone should train that day. The final word still needs a person paying attention, because attention is the part no sensor has figured out how to replicate.
If 2026 is the year coaching gets more automated, I think the honest response isn't resistance, it's clarity about which half of the job was never going to automate. Reading a number is easy now. Reading a person still isn't, and I don't think it ever will be.
