
Today is consecutive day 2,150 of my daily run, and day six of a 30 day experiment. I am taking a business concept from build to launch to delivery, and hopefully to profit, with AI carrying 90% of the workload. The other 10% is people pointing the AI in the right direction.
Another 50 day milestone sits behind me. I have run 16,513.28km so far, which leaves 23,561.72km of the 40,075km lap of the world. Daily running has taught me to value the unglamorous days, and today was one of them. Nothing dramatic happened. We had a planning meeting, and the lesson came from the structure of it.
Each week I meet the two other members of staff who give part of their time to this project. We agreed the roles clearly.
- I build the product. AI writes the modules. I supply the prompts and the briefs, and I direct the build rather than type the code. - One team member takes sales and marketing. This covers finishing the website, checking every integration works so people sign up and subscribe through it, building the content machine, and running paid ads. AI does the groundwork. The team member provides oversight. - The other team member takes testing, onboarding and support. They use AI to test, then stress test everything we build and try to break it before a customer does. They also shape the user experience and the customer support process.
We also made one significant business decision. We are moving to a subscription model. A one-off sale demands too much work for each customer. A decent product at a fair price, sold on subscription, builds recurring revenue, and recurring revenue matters.
We are ahead of where we planned to be on day six. I have built roughly two thirds of the product. This week my aim is to finish the core modules. One team member is working on the content machine, which takes one video and turns it into formats for each social platform, a blog and a podcast. The other is working through the web app we have built so far, using AI to test and troubleshoot. If the week goes to plan, we may have version one ready by the end of it. I will report what actually happens, not what I hope will happen.
Before this experiment I expected AI to remove the need for structure. Six days in, I believe the opposite. AI carries more of the work when someone has set it up properly first. Set up is the 10% that produces the 90%.
If you run a small business and want AI to take on repeatable work, these are the questions I would answer first.
- Who owns each area of the business? Name one person for each. - What does AI do in that area, and what does the human check? - What does a good result look like, written down in plain words? - Who tests the work before a customer sees it?
Without those answers, AI produces a lot of output with no direction. With them, the human role shrinks to briefing, reviewing and deciding. That is the 10%.
My streak works the same way. I do not decide each morning whether to run. The decision was made long ago. The system, which is running barefoot-style every day and filming it, removes the daily negotiation. Effort goes into the run, not into the choosing. A good business structure does the same job. Roles, briefs and checks are agreed once, so daily effort goes into the work.
Consistency over intensity applies here as well. I do not need a heroic week of building. I need a clear plan, steady progress, and honest reporting on day seven, day eight and every day after.
I run and vlog every day as part of a fundraiser. The aim is to complete 40,075km, a lap of the world, and raise £1M for children's causes, including Great Ormond Street and BBC Children in Need. The streak is planned to run for around 16.5 years. At day 2,150, I have 23,561.72km to go. The distance gets covered one day at a time, and I do not skip days.
The business experiment sits beside the mission. It is a test of how far a small team and AI can go with the right structure. Whatever the 30 days produce, I will share the results truthfully.
If you are using AI in your own business, tell me where you divide the work between AI and people. Drop a comment on the episode and I will reply to anything relevant. If you believe in the mission, subscribe, like and comment. The more people who see the vlog, the more we raise, and the more children we help.
Watch the full episode: https://youtu.be/39JrHtZ8F6c





