Skip to content

Mr Z

AI Prompt for Strength Programming: By a 19-Year Coach

Francesco Zen, founder of ZTraining · Published 8 July 2026 · Updated 8 July 2026

The prompt

Copy-paste prompt, built from 19 years of coaching
Act as my strength coach, not a template generator. Before programming anything, ask me for: my training age (how long I have trained consistently, not since I first touched a barbell), my current working weights on the main lifts I use, how many days a week I can train, session time available, any injuries or lifts I need to work around, and my main strength goal. Build my programming around the lifts and days I gave you. Do not add lifts I did not mention or days I do not have. Set my working weights from what I told you I currently lift, starting slightly below it, not from a percentage of a max I have not tested. Give me a clear week-to-week progression: add weight only when the current weight is steady, with more reps than last time, in full control and feeling easier; when adding weight stops being realistic, slow the lift down instead. Treat a stalled lift as information. When progress stops, tell me specifically what to change, and do not just tell me to push harder or be more disciplined. Programme recovery as part of the plan, not an afterthought: tell me what the minimum rest between heavy sessions of the same lift should be for my training age. If I miss a session, adjust from where I am. Do not make me repeat the whole week or deload unnecessarily. Keep assistance work in proportion. The main lifts drive the progress, assistance supports them, not the other way around. Be honest about the rate of progress someone at my training age can expect, and say so in numbers I can hold you to. Ask me the setup questions now, before writing any programming.

Why do generic strength programmes stop working?

Generic strength programmes fail at the two ends of the same problem: they either prescribe percentages of a one-rep max a person has never tested, or they progress on a fixed schedule that ignores what happened in the gym last week. This prompt anchors everything to reported working weights and forces the programming to state, in advance, what happens on the first failed session, which is the moment most templates simply have no answer for.

The stalled-lift rule matters more with every year of training age. A beginner progresses almost regardless of programming quality. Past that stage, the difference between a programme and a template is precisely what it does when the bar stops moving, and a plan with no written answer to that question was never really a plan.

A real-world example

The right output rebuilds the week around the three lifts with the deadlift treated as the diagnostic problem, a specific change to its volume or frequency, a defined check-in point, and working weights set from current reality rather than an aspirational max.

What this prompt is built to prevent is the default AI answer to a stall: add more volume everywhere and hope. More is not a diagnosis.

What can a strength prompt not see under the bar?

Strength programming is a feedback loop, and a prompt only sees one side of it. It prescribes, but it never learns what happened under the bar last Tuesday unless you type every session back in, every time, and even then it is reasoning from your description rather than a record it has kept.

It also cannot see bar speed, technique breakdown under load, or the difference between a hard set and a failed one, distinctions that decide whether the right call is pushing on or backing off. That judgment is exactly the part of strength coaching that does not compress into a single text answer.

Common questions

Can AI write a good strength programme?

It can write a solid one if you feed it real data: actual working weights, honest training age, and true availability. Its structural weakness is the feedback loop, it cannot see what happened in your last session unless you report it manually.

Why did my lifts stop progressing?

A stall is information, not a discipline problem. The usual suspects are recovery that no longer matches the training stress, too much added too fast, or a lift that needs its volume or frequency changed rather than simply pushed harder.

Should I train with percentages of my one-rep max?

Only if that max is real and recent. Working from an untested or outdated max is how people end up grinding weights that are wrong in both directions, so basing loads on current working weights is the safer anchor for most lifters.

Related: The complete AI fitness prompt library · AI prompt for your first training plan · Best AI prompt for a workout plan

Mr Z, the AI assistant built on his method, brings the same approach to a wider audience.

EREPS Level 6, strength and conditioning. About Francesco · How this guide was written

General information, not medical advice: if you have a health condition, or have not exercised for a long time, talk to your GP before you start.