Clone Looping System: The Idea Behind CLS – by Mario De Luca

Where the idea comes from

At 18 I tried four things: dropshipping, trading, memecoins, my own fashion label. Four times nothing, five-figure debt at the end, all of it from my childhood bedroom. I wrote about it in detail in my piece on those four failed attempts. The most important takeaway: I was always looking for the next tactic, never for a system.

The turning point came when I watched an AI build a website in minutes. It was instantly clear to me: this isn't a tool, it's an employee. I offered that first clone to businesses, and that's how my first clients came in. Out of that experience CLS later grew – an attempt to write down the path I took in a way others can follow.

What the name means

Clone, because it's about your own AI clone: a system that does work a person at a computer would otherwise do. Looping, because the process doesn't happen once, it repeats – build, deploy, support, improve, and then the same with the next business. System, because that's exactly what it's meant to be, not a collection of tricks.

The shorthand CTRL + L + $ sums it up: stay in control of your own clone, think of the process as a loop – and run the whole thing as a business, not a hobby. The dollar sign is a direction, not a number. How much anyone earns with it depends on so many things that any promise would be dishonest.

Why the car comparison

My favorite way to explain CLS is a car rental. The car belongs to you. The customer uses it for as long as they need it. You make sure it runs, you service it and you take care of it when something breaks. The customer doesn't need to know how the engine works – they just want to get where they're going.

That's how the clone is meant to work. A local business doesn't want to understand AI. It wants a task done that's costing it time. You build the clone, you look after it, and the business uses it. That's a different model from "I sell a project once and I'm gone" – it's an ongoing relationship. How to set up a service like that in general, I've described in Starting an AI agency (German).

What the process looks like, in principle

Without going into the details that live inside the community, CLS can be described in four steps:

1. Build the clone. With Claude and Claude Code you build a system that reliably handles one specific task. Not ten tasks halfway, but one task properly. What an agent like that can do in a business is in AI agents for business (German).
2. Find the right business. A local business with a problem the clone can solve. Not every business is a fit, and that's fine.
3. Have the conversation. This is where everything is decided, and no AI helps here. My scripts are deliberately linear, one thread from A to Z. I explain why in Sales: emotion beats systems.
4. Support and repeat. The clone runs, gets maintained, gets better. And then the process starts over with the next business. That's the looping.

What a clone like that costs to build and run is one of the most common questions. The honest answer depends on scope; how to calculate it properly is in What does an AI agent cost (German).

Not a course to consume

I've seen enough courses that you watch, tick off and forget. CLS is designed differently: as a process you execute. The community runs on Skool, and the entry point is claude-erfolg.de (German-language community). If you only watch there, you won't get much out of it. If you build, ask questions and take the next step, you get something you can actually put your hands on.

I'm not a guru, and I don't want to be one. I show what I build and use myself – my own operating system ESIS and my personal clone Jarvis are the examples from my own daily work. Output over theory.

What CLS explicitly is not

• No income promise. I don't name numbers anyone is going to hit. It wouldn't be honest, and in many places it wouldn't be allowed either.
• No shortcut. Building a clone, convincing a business and supporting it over time is work. AI takes some of it off your hands, not all of it.
• No legal advice. If you approach businesses, you need to know the rules on marketing and data protection. What I know from practice is in AI in sales (German). This is not legal advice; rules differ by country (in Germany: UWG/GDPR), and for a binding assessment you need a lawyer.

What I learned along the way

The first lesson: the clone is rarely the problem. Technically, a surprising amount can be built today. What things fail on is almost always the conversation, choosing the right business or the support afterwards. That's why sales sits at the center of CLS just as much as the tech.

The second: a system beats motivation. Motivation is there on good days and gone on bad ones. A process you simply run again works on the bad days too.

The third: honesty sells longer. If you promise what you can't deliver, you might win a client – and then lose them. I'd rather promise less and deliver more.

Why CLS is built on Claude

In CLS, everything is built with Claude by Anthropic and with Claude Code. That's not brand loyalty, it's experience: these are the tools I build with every day – my websites, my video engines that cut and caption short videos from raw footage, and my own operating system. I don't show anything in CLS that I don't use myself. Why I use Claude for longer, connected work is in Claude for business (German). Tools change; the principle stays: one clone, one task, one business, and then the next round.

Who CLS is for

CLS is for people who are willing to build something themselves and talk to real businesses. You don't have to be a computer scientist – I'm not one either. But you do have to be willing to try things, make mistakes and keep going. If you're looking for something that runs on its own, this isn't for you.

For companies that don't want to build their own clone but use a finished agent, there's NextCore. There we build and look after AI agents and automations for businesses.

I build all of this so that my father, who is still working at over 60, can stop someday. CLS is the part of it I share with others. More about my story is on About Mario De Luca.