All in on an AI coach

I’ll keep this simple.

When I first decided to run a marathon at the pace I wanted, I genuinely thought it would be straightforward. My original goal was to break 3:30. It sounded achievable.

It took me almost 18 months.

There were strong blocks, frustrating ones, lessons learned the hard way. Eventually I got there. But it wasn’t quick and it definitely wasn’t easy.

Now the next goal is 3 hours.

Not just because it’s a clean number, but because I want to qualify for the Boston Marathon. My age qualifier is 3:05, and typically you need to run about five minutes under that to be safe. Realistically, that means aiming at 3:00.

The challenge? There’s an August application deadline. Time matters.

So I did what most of us do in 2026 when we don’t know the answer.

I asked AI.

What surprised me wasn’t just that it gave me a plan — it was the quality of the conversation. It asked for context. It adjusted based on feedback. It wasn’t a generic marathon program; it felt structured and specific to me.

I was blown away.

So I made a decision: I’m going all in.

This project is me committing to follow a training plan fully guided by AI. No cherry-picking sessions. No doing my own thing when it suits me. If I’m testing this, I’m testing it properly.

And this blog is where I’ll document it — openly and transparently.

I’ll share:

• The transcripts of the chats.

• The weekly structure.

• Data from my runs and how AI analyses them.

• Adjustments along the way.

• What works and what doesn’t.

AI is only as good as the information you give it, so I’ll be feeding it everything — mileage, heart rate, pacing, fatigue, how I’m feeling mentally. The better the input, the better the output.

Already I’ve noticed how objective it is. It doesn’t care about ego. It just looks at the data and gives direction.

But there’s one thing it can’t fully do.

Hold me accountable.

Discipline is still on me. Showing up is still on me. Which makes this experiment bigger than just a marathon.

If AI tells me exactly what to do… and I don’t do it… and I miss the goal… was the plan the problem?

Or was I?

I’m genuinely curious to find out.

Can AI take me from 3:30 to sub-3?

Let’s see.