Dividing work when AI does 90% of the building is harder than I expected. Today is consecutive day 2,157 of my daily run, and day 11 of my 30 day AI challenge. The challenge is to take a business concept through build, launch and delivery, and hopefully into profit, within 30 working days. That is closer to six weeks than one month, and I prefer it that way.

Today was my weekly catch up with the team, and it gave me more clarity than any single build session has.

The problem I had been carrying

I had struggled to divide the labour. When AI is doing most of the repeatable work, the obvious question is what the people on the team should do. If I split the work badly, two people end up waiting on me, or doing tasks AI already does better.

I settled on this. I lead product development, because I know what the product should be. I have two team members, and each now has a clear lane.

The first has a sales and marketing background. Her lane is the marketing platform: the website, the sales funnels, the paid advertising campaigns, and the way we nurture leads until they book a call. She builds these systems with AI. I will take the sales calls myself in the early stages, so I stay close to what customers say.

The second has one purpose: quality control, quality assurance and user experience. Over time that grows into customer support, service and implementation, so the whole delivery side belongs to her. For now, she follows behind me as I vibe code. She finds the bugs, tests the fixes and keeps my work honest.

The two-tick rule

Our checking process is simple. There is a Slack channel for UX issues, and it is not for conversation. She lists each bug she finds. I fix it with AI and add a green tick. She retests it and, if it works, adds a second green tick.

One tick means I think it is fixed. Two ticks mean someone else has proved it. Only two ticks count as done. I like this rule because it removes opinion. Nobody argues about whether something works. The channel shows it.

The numbers for the next ten days

I have identified 29 modules I want built. I want them finished by the end of day 20, which means about three a day. Some are large. Some take a few minutes. On day 11 I hoped to complete a couple more before bed, and I will check that against the tally tomorrow.

The five days after day 20 are for migrating everything into a platform robust enough to handle up to 50 clients. My working target is 30 clients within 12 to 18 months. The offer is relatively high ticket, so a client count between 30 and 50 would put the business value in the region of half a million. I treat that as a barometer for a possible exit, not a promise. None of it is proven yet.

Why the weekly check-in matters

I also run an existing property business, and it needs time and focus. The weekly catch up covers both. It picks up what is going well in the property business and what we need to do to hit the launch deadline. Without the meeting, I would carry both lists in my head, and that is where things slip.

The lesson for me today is about sprint length. Most business goals use a 90 day sprint, which is standard. Shortening to 30 working days forces decisions. There is no room to leave problems for later, so regular communication matters more. A weekly check-in plus a light daily cadence catches pain points early, without micromanaging anyone. A small problem found in week two costs an afternoon. Found in week six, it costs the deadline.

What running has taught me about this

I have run every day for 2,157 days in barefoot-style footwear. The streak did not come from intensity. It came from a simple rule, repeated daily, with a way to notice problems early. A sore calf on day three is an adjustment. Ignored until day thirty, it ends a run of days.

The team structure works on the same principle. Clear lanes, a simple checking rule and a regular rhythm. Consistency over intensity applies to a project sprint as much as it does to a run.

Mission anchor

This vlog is part of a fundraiser for children's causes. I am attempting the ultimate ultra marathon: run vlogging on consecutive days until I have covered 40,075km, a lap of the world. So far I have covered 16,563.24km, which leaves 23,511.76km. My goal is to raise £1M for children's causes. The more people who watch, the more people hear about it, and that is why I post every day.

If you are dividing work between a team and AI, or running a short sprint, I would like to hear how you structure it. Leave a comment, and if you value the mission, follow, share or subscribe. I will be back tomorrow with day 12.

Watch the full episode: https://youtu.be/FpIys8tiomk