"Can you send us the project plan for the AI solution?"

An SME managing director asked me this last week. His team was inspired by the potential of AI. He wanted a predictable roadmap, something on safe ground. The question is all too understandable. It is rooted in the desire for security through planning.

An answer that I give more and more often:

"I don't know."

This is often followed by a frown. Because as an expert, I am expected to have a detailed plan. But my most valuable offer at the moment is usually: to show ways of dealing with uncertainty.

🎙️ You prefer to listen? Here's a short, AI-generated audio summary of this issue:

The trap of the grand plan

Why? Because in today's AI landscape, anyone who sells you a large, multi-year software project is doing you a disservice. The models from Google, OpenAI and the like are evolving on a weekly basis. We live in a VUCA world - volatile, uncertain, complex and ambiguous. Sticking to old, rigid planning methods is no longer "playing it safe", it is the biggest risk of allas well as Forbes recently aptly described.

Until now, it was perhaps still possible to somehow muddle through with long-term plans. But that is a thing of the past. Acceleration, especially in the AI sector, is so extreme today that rigid planning simply no longer works. The most successful approach is to surf agilely from one iteration to the next. Planning is replaced by experimentation. Instead of waiting for one big solution, we implement many small, quick partial solutions and learn continuously in the process.

So if a grand plan is the wrong way to go, what is the right first step?

The first step: Solving a specific problem

The answer is to start small and solve a real, pressing problem.

I know the feeling all too well. I recently had to write a detailed project report for a publicly funded initiative. A task I was downright dreading. A whole week of productive time that was missing for strategy and clients. That was my pain point. So I stopped being the "report writer" and became the "system builder". I created a customized AI assistant with Google Gemini, fed it with the relevant data and gave it the task of designing the report. The result: one week's work turned into three hours. The time we gained? We immediately invested it in planning a webinar for a new product - in other words, directly in sales-boosting work.

This example is not an isolated case. It is the result of a systematic method. A way of thinking that can be adapted from sport to the business world.

The system behind it: Putting an end to "junk miles"

I have borrowed a term from cycling for this type of work: "Junk miles". These are all the tedious everyday routines - business as usual - that leave us exhausted at the end of the day without having made any real progress. To break out of this hamster wheel, you need to consciously prioritize. It's about making three clear decisions:

1. What can an AI do better and faster? That's what we give away.

2. What needs our full concentration and critical thinking? We make time for this.

3. And everything in between? We put this to the test: Is it really still needed, or is it outdated and can be deleted without replacement?

In sport, this principle is called "Polarized Training". You spend most of your time at two extremes and eliminate the wasteful middle:

Zone 1: Automate radically (the "easy miles"). These are all the recurring, rule-based tasks that an AI can do better - as the example with the project report shows. This is the true leverage of AI: automating time-consuming, mostly unpopular tasks to free up time for high-value activities.

Zone 3: Focused deep work (the "hard miles"). This is the work that makes the difference: developing the corporate strategy, mentoring key people, creative brainstorming for a new service. This is where all the energy goes that you save in zone 1 through automation.

Eliminate Zone 2: The "Junk Miles". This is the gray middle range, the enemy of every peak performance. The meetings without a clear agenda, the reports "for the record". Here you have to radically question and shorten.

This change is more than just an efficiency hack. It is a fundamental mindset shift. It allows you to confidently say "I don't know" to a big, risky plan because you know a more flexible, focused, agile way.

The first step for you and your SME is therefore always the same: identify and eliminate your first "junk mile".

Theory is good, practice is better.

That's why I would like to invite you to an event. We are organizing with MMIND.ai our first AI Café: An afternoon full of inspiration for your SME.

• When: Wednesday, September 10, 2025, 4:30 - 7:00 p.m.

• Where: RUUF, Duxgass 55, 9494 Schaan, Liechtenstein

What you can expect: Together with the KMU-HSG, we are presenting the new "Digitalization Guide". But even more importantly: companies from Liechtenstein such as b_smart, sl.one, HSW AG, CSL Corporate Services and CFP Business Consulting show their successfully mastered digitization and AI projects and invite you to Exchange of experience from SME to SME in.

Places are limited. Register now here free of charge and mark the date in your calendar.

P.S. I have graphically summarized the essence of the new, polarized way of working. You can download the file here and use it as a guide for you and your team to find and eliminate your first "junk mile":