They Kept The Dashboard And Sacked The Coach

How you coach clients and use data is going to become a really interesting topic in the next few years, specifically in certain niches.

When I worked more in sport - obviously data is a big part of that. Elite athletes are monitored in a variety of ways to inform training plans, nutrition interventions, game time etc.

But the data never went straight to them. It passed through a coach, an S&C, a nutritionist - someone who read it, decided what mattered, and told the athlete only what they needed to know.

Consumer wearables are now taking that dashboard, deleting the interpreter (replacing with a very generic algorithm) and pointing it at everyone.

Now? AI is arriving as the interpreter. Cheap, always on, endlessly patient - and a lot of people are going to conclude they don’t need a coach at all.

That conclusion is going to feel obvious right up until it doesn’t work.

I will say that this approach will work for a lot of people - I collect a LOT of data now (although still use a coach, which tells you something)

Data anxiety is already here - It’s a predictable result of handing someone a readout with nobody to interpret it properly, understand the gaps in it, and push back on areas where it’s wrong.

We’re already seeing Orthosomnia - people sleeping worse from chasing a sleep score. CGM users frightened of a completely normal glucose response to a completely normal meal. We’ll see more of this, as tools and tech do the job of the coach, instead of just providing data for the coach to interpret.

AI doesn’t repair a person’s relationship with their data - it amplifies it. If the number already lands as a verdict rather than as information, then an AI that references your data conversationally, every day, in natural language, is not a helpful assistant.

It’s judgey. And if you work with people who are upset by their scale weight number? This will be worse, not better.

The person who feels judged by their smartwatch will self-select straight out of AI coaching.

So the category will disproportionately serve the already-data-comfortable - the spreadsheet dieters, the people who were succeeding anyway - and it will report excellent outcomes, because it is measuring a population pre-filtered for the exact trait that predicts success with or without the tool. The people who most need help never enter the funnel, so the results will look great and mean nothing.

I don’t think the thing for coaches to do now is to create loads of posts about why they’re better than AI - But to highlight what it can do well. The logging, the recall, the pattern-spotting across months no human could hold in their head. The performance of a coach does not depend on remembering the client’s snacks they had last Tuesday.

What can’t always be handed over is the interpreting layer, because for most behaviour change, developing the capacity to notice, interpret and choose IS the change. Offload that and the client is compliant while the tool is present and helpless the moment it isn’t.

The seat between the person and the number is still there. It’s just empty. Coaches who compete on knowledge, when knowledge has been commoditised - will struggle.

But if you can lean into what AI will do, and also act as the filter for clients?

You can have the seat back.