AI Strength Training Plan: Progressive Overload, Automated

An AI strength training plan earns that name by reading your logged sets instead of following a fixed schedule. A printed program picks your sets, reps, and weight progression for the next twelve weeks before you have lifted a single rep. It might tell you to add five pounds every week on squat regardless of whether last week's set moved fast and clean or ground to a near-failure grind. That works for a few weeks, then it stops matching reality, because a static template has no way to see how you actually performed.

Real progressive overload requires trend data across sessions, not a single number typed into a template once. You need to know whether your last three sets of an exercise came in above or below your typical rep speed, whether your estimated one-rep max is climbing or flat, and whether your volume for that muscle group this week is higher or lower than it was the week you last made progress. A spreadsheet cannot see any of that unless you build the tracking yourself.

Coach Jamie Reyes reads your full logged history, not just your last workout, before suggesting the next session's numbers. That means a lift you have been crushing for three sessions in a row gets pushed harder, while a lift where your reps have been slipping gets held steady or backed off, all without you having to notice the pattern yourself and manually adjust the plan.

Sets per muscle group per week is the single biggest lever in most strength and hypertrophy programming. Most lifters make consistent progress somewhere in the range of 10 to 20 hard sets per muscle group per week, spread across two or more sessions rather than crammed into one. Beginners tend to respond well nearer the low end of that range, while trained lifters often need the higher end just to keep the needle moving.

The catch is that almost nobody counts this by hand accurately. Chest gets hit directly on bench day and indirectly on overhead press and dip days; back gets credit from rows, pulldowns, and deadlift variations that only partially train it. Wellness Project tallies volume per muscle automatically from your logged workouts, rolling up every set that touches a given muscle group across the week, so it is visible the moment a muscle is sitting under its target range or has quietly crept past what your recovery can support.

That rollup is what makes the plan actionable instead of just informative. If shoulders have been under-trained for two weeks running, the next session's recommendation nudges volume up there specifically. If hamstrings have been climbing past a sustainable weekly total, the plan holds or trims rather than stacking another heavy session on top.

Every logged set gets compared against your all-time best for that exercise, measured by weight times reps and by estimated one-rep max, so a personal record is caught the instant it happens instead of relying on you to remember your old numbers or test a fresh max on a dedicated day. There is no separate max-testing session to schedule and no spreadsheet tab tracking best lifts by hand.

This ties directly into how the program progresses. A fresh PR is a signal, not just a badge on the dashboard: it tells the plan that your capacity on that lift has genuinely moved up, which is exactly the evidence progressive overload needs before it nudges the next session's target weight or rep target higher. Without that detection, a program either progresses blindly on a fixed schedule or lags behind your real capability by weeks.

None of this requires setting up a program in advance. The plan builds itself from the sets you log, and it keeps adjusting every time you train.

Progression that ignores recovery eventually breaks something, a stalled lift at best, a strain or an overuse injury at worst. Sleep, HRV, and the muscle fatigue heat map, an estimate of how ready each muscle group is to train again based on your logged sets, training intensity, and wearable signals, decaying over time as you recover, feed into whether the next session should push harder or pull back.

This is an estimate used to guide programming, not a diagnostic tool and not a substitute for how a joint or muscle actually feels. It draws on whichever wearable you have connected, Apple Health, Oura, Fitbit, or Health Connect devices like Garmin and Whoop, so the recovery side of the equation updates automatically as your sleep and HRV data comes in, the same way your training volume updates automatically as you log sets.

In practice this looks like a deload that shows up on its own: a week of short sleep and elevated resting heart rate softens the next session's targets before fatigue turns into a missed rep or a tweak, rather than after.