Give or take a few days, my friend opened his so-called “boutique” gym around this time last year.
The positioning was simple but bold: hyper-personalised personal training.
At first, I honestly thought he’d lost his mind.
The Setup Looked Insane (At First)
He poured a lot of money into:
- Matte black walls
- High-end equipment
- Big frameless mirrors
- Motivational quotes everywhere
It looked expensive.
And for the first couple of months, there wasn’t much traction at all.
I didn’t see how this would work.
The Pricing (Where It Gets Wild)
Fast forward to now.
He charges £80 for a 90-minute personal training session.
Here’s the kicker:
- His trainers aren’t highly certified professionals
- They’re mostly fit-looking university students
- He pays them around £15 per hour
On paper, that sounds insane.
The ChatGPT Layer (The Real Business Model)
Before a client’s first session, they fill out a detailed questionnaire on an iPad:
- Goals
- Injuries
- Experience level
- Lifestyle info
They even take a few “before” photos.
All of that data gets uploaded straight into ChatGPT.
From there, ChatGPT generates:
- A 12-week progressive training program
- Exercise form cues
- Motivational talking points tailored to the client
The trainers then:
- Pull up the program on their phone
- Guide the client through the session
- Literally read the ChatGPT output word for word
No improvisation.
No deep coaching experience required.
The Numbers (Rough Breakdown)
Here’s how it looks right now:
- £80 per session
- Around 4 sessions per day
- Across 5 trainers
- Average client books 2 sessions per week
- Most commit for at least 3 months
That puts monthly revenue at roughly £40,000.
Costs:
- Trainer wages: ~£11,000
- Gym overhead: ~£5,000
Leaving around £24,000 before taxes, leases, and misc expenses.
It’s not pure profit — but it’s still significant.
Why Clients Aren’t Complaining
So far, nobody’s complained.
My friend’s logic is simple:
“ChatGPT has access to thousands of peer-reviewed studies and decades of training data. It’s qualified enough.”
Clients feel:
- The program is personalised
- Sessions are structured
- Progress feels intentional
And honestly… most people wouldn’t know the difference.
The Scaling Question
This is where it gets interesting.
He now wants my help scaling this to two more locations.
I’m skeptical.
I don’t think many places can replicate:
- Central London foot traffic
- High-income density
- Willingness to pay £80/session
He argues the opposite — that many cities could support this model.
House of Cards or Scalable Business?
That’s the real question.
On one hand:
- Strong margins
- Systemised delivery
- Trainers are replaceable
- AI handles the “expertise”
On the other:
- Regulatory risk
- Liability concerns
- Dependence on AI outputs
- Brand trust at scale
I’m torn.
Is this a genuinely smart, scalable business model — or am I helping him build a house of cards?
Curious what others think.