Kodak didn’t fail because it couldn’t build a digital camera. It failed because it couldn’t accept that its actual product had changed. For decades, Kodak’s real business was the chemical film process, the thing sitting between the camera and the photograph. When that layer became redundant, the company that had once dominated its industry struggled to survive.
Personal training has a layer sitting between the client and the result too. It’s called the programme. And that layer is becoming redundant faster than most of the industry wants to admit.
The Programme Was Never Really the Product
Ask most certified personal trainers what they sell and they’ll describe some version of structured exercise programming: periodised training blocks, progressive overload, movement screening, macros, and recovery protocols. The curriculum that underpins most major certifications reflects this. NASM, ACE, ISSA and their equivalents are thorough, science-based, and built around a core assumption: the trainer’s primary value is their ability to design and deliver an effective training programme.
The uncomfortable truth is that this has never been the primary reason clients stay or leave.
Research indicates that nearly half of individuals who begin an exercise programme drop out within the first six months. The programme design almost never explains why.
Most coaching challenges aren’t about exercise selection. They’re about what happens in a client’s head before, during, and after a session. Adherence research points consistently to psychological need satisfaction, not training quality, as the deciding factor. Clients quit when they feel unseen, when goals feel arbitrary, when the emotional contract breaks down. They stay when they feel known.
The programme was always a proxy for the relationship. AI is about to make that very difficult to ignore.
What AI Actually Does Well
The AI personal trainer market was valued at over $16 billion in 2025 and is projected to reach $65 billion by 2033. That is not a niche development. That is a structural shift in who builds training programmes and at what cost.
AI personal trainer solutions now use adaptive learning frameworks and real-time feedback, resulting in personalised workout recommendations that evolve with user progress. Virtuagym’s MAX AI Coach can generate a personalised programme in 20 seconds using a library of over 6,000 exercises and individual client data including fitness level, goals, available equipment, and injury history. Peloton IQ uses computer vision to deliver form feedback, rep counting, and performance insights in real time. These are not approximations of coaching. They are, in the narrow domain of programming and form correction, credible alternatives to it.
Personal trainers who use AI tools can handle 30% more clients while maintaining quality, and AI-enhanced coaching services deliver up to 221% ROI. The efficiency gains are real. The risk is that efficiency and value are not the same thing, and the industry conflates them at its peril.
What AI does well is volume, consistency, and pattern recognition across data. It will write your client’s next four training blocks without fatigue or bias, adjust load based on HRV data without ego, and remember every metric from every session with perfect fidelity.
For a significant portion of what a mediocre trainer does in a session, AI is already competitive. For what a great trainer does, it is nowhere close.
The Illusion of Personalisation
The word “personalised” appears in almost every personal training marketing pitch ever written. It implies that the programme a client receives has been specifically designed around their unique physiology, psychology, goals, and circumstances. In most cases, this is a polite fiction.
The honest reality is that a large proportion of personal training involves relatively standardised templates applied with modest individual variation. A beginner gets a push/pull/legs split. A fat loss client gets a calorie deficit and three sessions of resistance training per week. A sports performance client gets periodised blocks with sport-specific conditioning. These are not bad programmes. They are evidence-based defaults dressed up as bespoke solutions.
AI does not just replicate this, it exposes it. When a client can download an app that asks them eight questions and produces a twelve-week programme with progressive overload, form cues, and automatic adjustment based on their wearable data, the question they will inevitably ask their trainer is: what exactly are you adding that this can’t?
Trainers who learn to use AI well will replace those who do not. But the more important point is that the trainers who survive won’t be the ones who learn to use AI as a programme-building tool. They’ll be the ones who stop pretending the programme was ever the point.
The Skill Set That Actually Matters
Behaviour change is the domain where AI is structurally weakest and where the best human coaches have always been structurally strongest.
The gap between the two is not closing quickly.
Research grounded in Self-Determination Theory shows that when people feel psychologically supported, exercise adherence improves significantly, while need frustration is linked with higher dropout and negative emotional responses. The three psychological needs at the centre of this theory, competence, autonomy, and relatedness, are all delivered primarily through human interaction. Feeling capable, feeling in control of your own journey, and feeling genuinely known by the person coaching you are not problems an algorithm solves well.
Motivational interviewing, the evidence-based counselling technique that underpins effective behaviour change coaching, is built around something AI cannot currently replicate: genuine attunement to ambivalence. A skilled coach using motivational interviewing doesn’t tell a resistant client what to do. They help the client hear themselves say why they want to change, in their own words, and then they get out of the way. That requires presence, timing, and the ability to sit with uncertainty rather than fill it with information. These are deeply human skills.
The trainer who recognises that a client is showing up burnt out and needs a conversation more than a conditioning session, who can read the room well enough to know that today the session needs to be half the planned intensity, who keeps a client in the gym on the worst weeks not through programming but through relationship: that trainer is not replaceable. The trainer who shows up with a clipboard and a template is.
The Profession’s Awkward Adolescence
The personal training industry is large, growing, and deeply uncomfortable with this moment. The US Bureau of Labor Statistics projects employment of fitness trainers and instructors to grow 12% between 2024 and 2034, far outpacing the 3.1% growth expected across all occupations. The headline numbers look healthy. Beneath them, the structural question is sharpening: as AI commoditises programme design, what exactly is the certified trainer’s differentiating value?
The certification industry has been slow to answer this. The dominant frameworks still weight exercise science heavily and behaviour change psychology lightly. A trainer can emerge from most certification programmes knowing the difference between concentric and eccentric loading, the mechanisms of metabolic adaptation, and the principles of progressive overload, without having spent meaningful time on motivational interviewing, psychological safety, or the neuroscience of habit formation.
These aren’t soft skills. They are the skills that determine whether a client returns for a second month.
The trainers building durable businesses in this environment share a common profile. They are less coach and more trusted advisor: the person a client calls when their relationship with food unravels, when a health scare changes everything, when they need someone who knows their history and can help them think clearly. Programming is one small part of that relationship. It was always the smallest part.
The Kodak Lesson
Kodak’s actual mistake wasn’t missing the digital camera. It was misidentifying what business it was in. The company thought it was in the film business. It was actually in the memory business: helping people capture and keep meaningful moments. Digital photography served that need better, and Kodak never made the conceptual leap in time.
Personal training is in the same position. The profession has misidentified itself as being in the programming business. It is actually in the behaviour change business: helping people build a consistent, sustainable relationship with their own health. AI serves the programming part of that better and cheaper than most human trainers do. It does not come close to serving the rest.
The trainers who understand this distinction will find that AI makes them significantly more powerful: freed from the administrative labour of programme design, they can invest fully in the thing that actually drives retention, results, and referrals. The trainers who don’t will find themselves competing on price with software that never has a bad day, never runs late, and never forgets a client’s name.