By Francesco Zen, the coach behind Mr Z Updated
Quick answer
A strong AI prompt for muscle gain anchors protein and calorie surplus to lean body mass, not total bodyweight, and asks how many days you can actually train before writing a single set. A modest surplus builds the same muscle as an aggressive one, with far less fat gained alongside it.
Copy-paste prompt, built from 19 years of coaching
Most generic muscle-gain prompts hand back a flat surplus and a bodybuilder split regardless of who is asking. This one starts from lean mass and real training days instead, since a beginner at 25% body fat and an advanced lifter at 12% need meaningfully different protein and surplus sizes even at the same scale weight, and a flat formula cannot tell them apart.
The surplus is deliberately kept modest. The rate of muscle protein synthesis is the ceiling on how fast muscle is actually built, and eating far above it does not raise that ceiling, it just stores the excess as fat. The most common muscle-gain mistake is not undereating, it is bulking into a body composition that then takes months to undo.
Take a 35-year-old consultant with four available training days and a bench and dumbbells at home. His lean mass sits well below his scale weight, so his protein target is set from lean mass and matched to structured training, not to what an advanced lifter would need.
The equipment question matters as much as the numbers: a plan built around machines he does not have is a plan that ends in week one, so the split is built as four home sessions with progression written in, not a six-day gym programme he cannot attend.
The target it gives you on day one stays frozen. When you have gained a few kilos and your lean mass has shifted, nothing updates unless you recalculate and paste the entire prompt again. It also cannot see whether your lifts are actually progressing, the one signal that says whether any of this is working, unless you report every session back yourself.
A prompt gives you a reasonable starting point. What it cannot do is notice the plan stalled in week five and change it, before you have spent three months maintaining instead of building.
Mr Z is an AI coach built on Francesco Zen's 19 years of coaching. It remembers your goals, your schedule and your progress, and adjusts the plan when your week changes: live form feedback from your camera, training that reshapes itself around a missed session, and nutrition you log by talking.
It can build a solid starting structure if you force it to gather real numbers first: training experience, equipment, and lean mass. Without that, you get a generic bodybuilder split that ignores your actual week.
Modest. A large surplus does not speed up muscle growth, it mostly adds fat, so the target should be enough to fuel training and recovery, not a big jump above maintenance.
Anchor it to lean body mass rather than total bodyweight, and match it to how you actually train. Recalculate it as your body changes, rather than keeping the number you started with.