When ChatGPT arrived, everyone became an AI expert overnight. Why I still rebuilt SIFAMO completely around AI, what separates cognitive transformation from digitalization, and how an AI system in our recruiting meant that only two employees left us in five years.
At the SIFAMO AI Summit 2025 in the Burj Khalifa, Tanishka from The Technology Express asked me what the real driving force behind SIFAMO was. This is the first part of the interview. It is about why a German software company decides to stop just building software and start making its systems intelligent, and why we went to Dubai to do it.
Part 1 of the interview with Tanishka (The Technology Express), SIFAMO AI Summit 2025, Burj Khalifa, Dubai.
Data instead of volume: why SIFAMO came into being
We are a German company and did classic software development for years. Along the way we kept seeing the same thing: companies digitalize their processes, but the processes stay slow. Instead of making the systems better, they hire more people. By the time ChatGPT arrived at the latest, it was clear to me where things were heading. The future has a direction: make systems more intelligent, automate more.
The personal reason was a different one. I hated working in companies where decisions were based purely on emotion. You give everything, you work hard, but nobody recognizes the facts because you are not the loudest person in the room. I wanted a company that decides based on data and facts, where the truly good employees become visible and not the loudest ones. That was the moment I said: we are going in the direction of AI.
From Berlin and Frankfurt to Dubai
In Germany we have two locations, Berlin and Frankfurt. AI is a topic there, but it is not at its peak, there are simply other problems to deal with. So we said: we have to go where the future is already being lived. For me, Dubai is the capital of technology. The people there have a different attitude. They want to know, they want to talk, they want to network and understand how you grow. That is exactly why we held our first own summit there.
Cognitive transformation instead of digitalization
We had the era of digitalization. Ten or twenty years ago we handled processes manually, then they became digital. For some companies that process is still running, but most have long been digitalized. What they want now is something different: to make their systems smart. I call this cognitive transformation. Take the digitalized systems and make them intelligent. That is the step that is due now, and it will change an enormous amount.
AI in recruiting: 90 percent match and two departures in five years
The simplest example is our own company. We had an HR department, we hired, and sometimes we hired the wrong people. The soft skills were good, the technical skills were missing. The reason was simple: developer CVs are almost impossible for regular HR staff to read. If it says Angular, HR says: no HTML, no CSS, rejection. Every developer knows that Angular is HTML and CSS.
Intelligent matching. A system checks whether the hard skills fit and also recognizes the skills that are indirectly contained in the frameworks mentioned.
Targeted interviews. At around 90 percent match, the system suggests the interview questions that verify the remaining 10 percent.
Humans for the soft skills. HR focuses on what HR is really good at: checking whether the person fits into the team.
The result: in five years, only two people have left our company. That is what we stand for. We take care of our employees, we want the best, and whoever works with us gets the best. How to build AI systems of this kind so that they hold up in regulated industries is something I described in my post on agentic AI and AI governance (German).
The ChatGPT moment: two years to an AI-powered ERP
We had already been thinking about it in 2021 and 2022, as Covid was coming to an end. Building web applications was nice, but everyone was doing it, and that could not be the future. When ChatGPT practically crashed the market and suddenly everyone was the best AI expert, we decided: we have to get on this boat, and properly. We started making all of our systems intelligent.
One week before the summit we completed our ERP system with the new AI board. That took two years. In return we can say today: almost everything we do runs with AI. At the end of the conversation came the question of how this fits together with clients like Bosch and Zurich and the strict European data protection rules. The answer, namely how we train AI without using personal data (German), is the second part of the interview.
What about you: are you still in the middle of digitalization, or already asking how your systems become intelligent? Let me know on LinkedIn or under the video on my YouTube channel.
Frequently asked questions
What is cognitive transformation?
The step after digitalization: processes and systems that are already digital are made intelligent with AI, so that they prepare decisions, recognize patterns and take over routine work instead of just storing and displaying data.
Does AI in recruiting replace the HR department?
No. In our setup, the system takes over what HR finds difficult, namely evaluating technical CVs. The decision on soft skills and team fit stays with people. The combination has reduced our turnover to two departures in five years.
What makes someone an AI expert today?
Not the title, but implemented systems that run in everyday operations. Anyone who has rebuilt their own company around AI, including ERP and recruiting, knows where the stumbling blocks are. Everything else is opinion.
When ChatGPT arrived, everyone became an AI expert overnight. Why I still rebuilt SIFAMO completely around AI, what separates cognitive transformation from digitalization, and how an AI system in our recruiting meant that only two employees left us in five years.
At the SIFAMO AI Summit 2025 in the Burj Khalifa, Tanishka from The Technology Express asked me what the real driving force behind SIFAMO was. This is the first part of the interview. It is about why a German software company decides to stop just building software and start making its systems intelligent, and why we went to Dubai to do it.
Data instead of volume: why SIFAMO came into being
We are a German company and did classic software development for years. Along the way we kept seeing the same thing: companies digitalize their processes, but the processes stay slow. Instead of making the systems better, they hire more people. By the time ChatGPT arrived at the latest, it was clear to me where things were heading. The future has a direction: make systems more intelligent, automate more.
The personal reason was a different one. I hated working in companies where decisions were based purely on emotion. You give everything, you work hard, but nobody recognizes the facts because you are not the loudest person in the room. I wanted a company that decides based on data and facts, where the truly good employees become visible and not the loudest ones. That was the moment I said: we are going in the direction of AI.
From Berlin and Frankfurt to Dubai
In Germany we have two locations, Berlin and Frankfurt. AI is a topic there, but it is not at its peak, there are simply other problems to deal with. So we said: we have to go where the future is already being lived. For me, Dubai is the capital of technology. The people there have a different attitude. They want to know, they want to talk, they want to network and understand how you grow. That is exactly why we held our first own summit there.
Cognitive transformation instead of digitalization
We had the era of digitalization. Ten or twenty years ago we handled processes manually, then they became digital. For some companies that process is still running, but most have long been digitalized. What they want now is something different: to make their systems smart. I call this cognitive transformation. Take the digitalized systems and make them intelligent. That is the step that is due now, and it will change an enormous amount.
AI in recruiting: 90 percent match and two departures in five years
The simplest example is our own company. We had an HR department, we hired, and sometimes we hired the wrong people. The soft skills were good, the technical skills were missing. The reason was simple: developer CVs are almost impossible for regular HR staff to read. If it says Angular, HR says: no HTML, no CSS, rejection. Every developer knows that Angular is HTML and CSS.
The result: in five years, only two people have left our company. That is what we stand for. We take care of our employees, we want the best, and whoever works with us gets the best. How to build AI systems of this kind so that they hold up in regulated industries is something I described in my post on agentic AI and AI governance (German).
The ChatGPT moment: two years to an AI-powered ERP
We had already been thinking about it in 2021 and 2022, as Covid was coming to an end. Building web applications was nice, but everyone was doing it, and that could not be the future. When ChatGPT practically crashed the market and suddenly everyone was the best AI expert, we decided: we have to get on this boat, and properly. We started making all of our systems intelligent.
One week before the summit we completed our ERP system with the new AI board. That took two years. In return we can say today: almost everything we do runs with AI. At the end of the conversation came the question of how this fits together with clients like Bosch and Zurich and the strict European data protection rules. The answer, namely how we train AI without using personal data (German), is the second part of the interview.
What about you: are you still in the middle of digitalization, or already asking how your systems become intelligent? Let me know on LinkedIn or under the video on my YouTube channel.
Frequently asked questions
What is cognitive transformation?
The step after digitalization: processes and systems that are already digital are made intelligent with AI, so that they prepare decisions, recognize patterns and take over routine work instead of just storing and displaying data.
Does AI in recruiting replace the HR department?
No. In our setup, the system takes over what HR finds difficult, namely evaluating technical CVs. The decision on soft skills and team fit stays with people. The combination has reduced our turnover to two departures in five years.
What makes someone an AI expert today?
Not the title, but implemented systems that run in everyday operations. Anyone who has rebuilt their own company around AI, including ERP and recruiting, knows where the stumbling blocks are. Everything else is opinion.
Related: AI Consulting · Umair Zaffar as AI Expert
Recent Posts
Recent Comments