The debate about whether AI will replace personal trainers is the wrong question. The right one is: where does AI save time without compromising quality, and where does it introduce risk for the client? This article offers an honest framework — no tech scepticism, no hype.

Three areas where AI excels

Initial programme generation

Given a structured profile — goal, level, equipment, medical constraints, target frequency — a model can produce a coherent programme outline in seconds: movement pattern coverage, volume dosing, warm-up / main work / accessory structure. The time saved on client onboarding is real. The outline still needs reviewing and adjusting, exactly as you would with a junior assistant.

Generic meal plan

Basal metabolic rate calculation (Mifflin-St Jeor formula), macronutrient distribution, suggested meal structure. For a coach working strictly within their legal scope — general nutritional guidance, not clinical treatment — this is a valuable educational support tool. See our article on the scope of nutritional coaching.

Progress analysis and adaptation

Automatic detection of load plateaus, gaps between planned and completed sessions, and patterns of missed workouts. AI surfaces the signal from the noise; the coach decides how to respond.

Areas where AI still falls short

What AI does well

Generating structured outputs from defined parameters
Calculations (calories, volumes, 1RM projections)
Detecting patterns in tracking data
Rephrasing technical concepts in plain language

What AI does poorly

Reading between the lines in a consultation
Assessing movement quality by eye
Decision-making in unusual clinical contexts
Building a relationship of trust

An AI that recommends a programme without knowing the client's state on a given day — fatigue, pain, work stress — produces a recommendation that is disconnected from reality. The coach remains essential as a contextual filter.

Three principles for using AI properly

What separates useful AI usage from lazy AI usage comes down to three ground rules.

AI proposes, the coach validates — no AI-generated output is sent to a client without manual review and personalisation.
Sensitive data stays under control — the software provider must document which data is sent to the model, which third-party provider is used, and how it is stored.
The client knows AI is involved — transparency is now an expectation. It does not diminish perceived value; on the contrary, framing the offering as "AI + coach expertise" makes your positioning more professional.

The regulatory landscape: what's coming

The EU AI Act, being phased in progressively, classifies use cases by risk level. Fitness coaching tools fall mostly into the limited-risk category, but the coach remains responsible for decisions made on the basis of AI recommendations. Traceability of those decisions is becoming a compliance matter.

The right reflex day-to-day

Rather than using AI to produce a deliverable "instead of you", use it to accelerate a deliverable you would have produced anyway. The initial programme the AI generates is a first draft; the final version, after your adjustments, carries your signature. The automatically generated monthly report is raw material; the conversation with the client during the session is where the real value is delivered.

AI augments the coach — it doesn't replace them

The coaches who get the most out of it are those who reinvest the time saved into deepening client relationships, not those who simply pocket it as extra margin.

Pour aller plus loin

Ready to transform your coaching?

Stay In Shape gives you all the tools to manage, track and retain your clients — 14-day free trial, no credit card.

Try for free