How did you get from 80 hours a week to 30?

I built an AI co-pilot on top of my own operational stack and plugged it into real data. It reads the ERP, watches 8 inboxes, and tracks 10 departments. The work that used to eat my evenings now runs without me. The 50 hours I got back go to learning, family, and spotting the next opening.

How long does a build take?

I built the first version of my own system in 60 days. Ugly, iterative work on real accounts, not a demo. For a client the timeline depends on how messy the data is, but 60 days is the honest benchmark, not a sales number.

Do you coach, or do you build it for me?

Both, and you pick. In 1-on-1 coaching we work on your real problems and something gets built or decided every session. If you would rather have it done for you, I build the custom infrastructure directly. Most people start with coaching and move to a build once they see what is possible.

What does the AI actually do day to day?

It monitors 8 inboxes and routes what matters. It reads the ERP live through a custom API. It writes meeting summaries, tracks staff and tasks across 10 departments, and pings me through the day instead of waiting to be asked. I open one dashboard with my morning coffee and see every department at a glance.

Will this work for my company, or only for food wholesale?

The pattern travels. I run Ratouli Foods, a 14-year food business, but the method is the same anywhere one person holds too much in their head. If your prices, suppliers, and decisions live between your ears, this is built for exactly that.

Message me on WhatsApp for a free 15-minute intro. I show you one AI hack you can use that same day. If it clicks, we take it further.

Did revenue drop when you cut to 30 hours?

No. The co-pilot did not replace the decisions. It replaced the hours I spent gathering the information to make them. The decisions still get made, faster, with the same person behind them.

Could someone do this without being technical?

The building part took technical work, and I will not pretend otherwise. But the hard part was not the code. It was being willing to write down how the business actually runs, and only the owner can do that.

Why 30 hours and not less?

Because 30 keeps me in contact with reality. A co-pilot you no longer understand is not an advantage, it is a risk you cannot see. Thirty hours is the line where I still know what good looks like.

Writing

1 July 20264 min read

Why I cut my company to 30 hours a week

For fourteen years I ran Ratouli Foods at 80 hours a week, with the whole company living in my head. In 60 days I built an AI co-pilot on the live accounts and brought my week down to 30 hours, without dropping revenue.

For fourteen years I ran Ratouli Foods at something like 80 hours a week. Orders, margins, supplier calls, customs paperwork, the production schedule. All of it lived in one place: my head. When a container ran late or a price moved, people called me, because I was the only system that knew everything.

I used to be proud of that. Eighty hours felt like proof. If everything ran through me, then I mattered, and the long days were the receipt. That is the trap, and I want to name it plainly. Busy felt like indispensable. I had mistaken the volume of my hours for the value of my work, and that mistake kept me stuck for years.

Being needed is a comfortable place to hide. As long as I was the bottleneck, I never had to ask whether the bottleneck should exist.

The 60 days

In February 2026 I gave myself two months to build something different. An AI co-pilot, sitting on top of the real Ratouli Foods accounts. Not a demo. Not a slide. I plugged it into the live ERP, the actual inboxes, the real margins per product line.

The work was ugly. I would wire up one piece, watch it misread a supplier email, fix that, then find it had the wrong cost base on a product and start over. Most days felt like going backwards. Honestly, the first three weeks produced almost nothing I trusted. I kept going because the alternative was another fourteen years of being the only copy of the company.

By the end of the 60 days I had a co-pilot that held the context with me. It read the inboxes, flagged the margins that moved, drafted the replies I used to write at 11pm. My weekly hours went from around 80 to about 30.

What 50 hours actually buys

Those 50 reclaimed hours were the whole point, not a side effect. So let me be specific about where they went. Some go to learning AI deeper, because the co-pilot is only as good as the person teaching it. Some go to spotting market openings I used to be too tired to see. A real chunk goes to friends and family, the part of my life that fourteen years of 80-hour weeks had quietly underpriced.

The dashboard is the tool. It is not the trophy. I do not care about a screenshot that looks impressive. I care about the Tuesday afternoon I now spend away from the desk.

The honest cost

This was not clean. Things broke. The co-pilot sent me down the wrong pricing path twice before I learned to check its math instead of trusting it. I had to re-learn parts of my own business that had gone on autopilot, because you cannot hand context to a system you can no longer explain yourself. Writing down how I actually made decisions turned out to be harder than making them.

And 30 hours is a floor I chose. It is not a number I bragged my way down to, and it is not heading to zero. At 30 hours I am still close enough to the work to feel when something is off. Below that I would lose the instinct that makes the co-pilot worth supervising. The point was never to disappear. It was to stop being the single point of failure.

Here is the principle I would give anyone running their own thing. Cutting your hours only works if you first move the context out of your skull and into a system that holds it with you. Cut the hours without doing that and you have not freed yourself. You have just stopped answering the phone.

Questions I get asked

Did revenue drop when you cut to 30 hours?: No. The co-pilot did not replace the decisions. It replaced the hours I spent gathering the information to make them. The decisions still get made, faster, with the same person behind them.
Could someone do this without being technical?: The building part took technical work, and I will not pretend otherwise. But the hard part was not the code. It was being willing to write down how the business actually runs, and only the owner can do that.
Why 30 hours and not less?: Because 30 keeps me in contact with reality. A co-pilot you no longer understand is not an advantage, it is a risk you cannot see. Thirty hours is the line where I still know what good looks like.
How long did it take to build the co-pilot?: Two months. I started in February 2026 and gave myself a 60-day window. The first three weeks produced almost nothing I trusted. The working version came together over the back half of that period.

All writing