It helps, but I believe it's also in the playbook. One of the essential differences between American and German or European marketing is the difference between 'Principles First' and 'Applications First'. 'Principles First' is about how something works, what the principles are, and 'Applications First' is about what the product brings you. For example, if I were marketing a miracle cure, in Germany I might say it has great minerals and stimulates circulation, while in the US I'd say you'll be healthier and lose weight. At Celonis, we've adjusted for this too. We ran a campaign in the Wall Street Journal with a '10 Million Dollar Challenge'—a full-page ad saying we'd save you 10 million, and if we didn't, we'd donate 100,000 to charity. It was a huge success in the US, but in Germany, people weren't interested. In Germany, the pitch is always about how the technology works, while in the US, it's about the use cases and what competitors have done. It's a different mindset, so you have to adapt your marketing. However, our brand as a German company with three German founders also works well in the US, as people trust German engineering and complexity. You don't have to completely forget your roots, but you do need to change how you communicate.
Now you have to tell us, how often were you right with the challenge and how much ended up in the donation account?
We didn't lose a single one. You didn't have to pay anything into the donation account.
Okay. You also have customers in Asia today, you just mentioned India. Are there other examples? What are you learning about this huge market? It's really one of the fastest growing.
It depends on where in Asia. We do less in China. We focus on Japan, Southeast Asia, and of course India. What we see there is that they are incredibly innovative and hungry. In India, it's incredible how it has developed. The times of it being just an 'extended workbench' are over. You have incredibly smart, well-educated people there. The IITs don't have to hide from any American or European university in terms of how hard it is to get in and the quality of education. There's a lot of innovation power, and you can see it's now the second or third generation with good education. The society is maturing. Of course, there are still many people in poverty, but a lot has moved. Japan is completely different. What's interesting in Japan is that process optimization is deeply ingrained in the culture.
What does that mean for you concretely?
The Japanese are very detail-oriented when it comes to marketing. It's not about how it works, but the details of how it works. I remember visiting the CEO of a large Japanese company. I had my briefing, prepared well, but he wanted to spend 45 minutes just talking about how our data model works. He asked questions where I thought, wow, he really gets it. We ended up talking for an hour about technical details. It's very different there. They are very detail-oriented and want to optimize everything. If you've been to Tokyo, everything works perfectly. We had some logistical challenges at the conference today, and that wouldn't happen in Japan. It fits well with what Celonis does—perfecting processes. Every country is different. What's clear in Asia is the huge innovation drive. But I'm an optimist, and even in Europe, this AI topic and what's happening globally is a wake-up call. We see it in Europe too. Today at this tech conference, there were probably 50 or 100 CEOs from large companies like Sanofi, Valeo—big firms. It's different than 5 years ago.
Yes, absolutely. And all these companies, whether from Japan, India, the US, France, or Germany, face the same challenge. On one hand, we have a technology that promises miracles. On the other, AI pilot projects rarely bring a return on investment. The money invested doesn't generate value. What's the short analysis? Why is that?
I believe there are two essential factors. First, AI is trained on publicly accessible data. For example, ChatGPT is trained on a petabyte of data. JPMorgan, for instance, has 150 petabytes of data. You have a lot of data in your company that AI models can't access. Second, the organization of data and processes leaves much to be desired. It's not unusual for a process like order-to-cash to run across five or six different IT systems, with half the work happening in Excel sheets and emails. If you try to put a GPT on such complex processes, it doesn't help much. That's why you need a digital twin, process intelligence, to really clean up and reinvent the process. I can give examples, but this is something companies must do to be successful.
The conference is called 'Adopt AI'. Most people first think of ChatGPT and similar tools that employees need to adopt. Is that the wrong approach, focusing too much on individual use?
I think there are different dimensions. One is giving all employees access to ChatGPT and feeding it with company data. That's certainly useful—we call it 'Knowledge, Search and Retrieval' to make knowledge more accessible. We do that at Celonis, and it's helpful. Am I more efficient? I don't know, but it's good. The second is code generation, where developers generate parts of code with AI. The third is conversations, like customer conversations. These three areas are relatively quick to implement and have value, but they're not transformative. You won't save hundreds of millions that way. For that, you need to bring AI into the core processes—how I pay invoices, manage orders, create offers, plan production. For example, a large bank customer used to take 40 days to onboard a new business customer; now it's under a week. You can't just give employees ChatGPT; you have to rethink the process. First, analyze it to understand where time is wasted. Then redesign it—what can AI take over, what must a human do, what can be fully automated? Then implement and monitor it. We call it 'Analyze, Design, Operate'. If you don't reinvent these processes, you miss huge value. Many are still at the beginning.
I don't want to misunderstand—the possibilities with generative AI behind ChatGPT are limited. You yourself recalled at Celosphere how it was when ChatGPT launched 3 years ago, and you spent Christmas exploring it and getting your colleagues excited, saying everything would change. Can you give a concrete customer example of where the leverage of generative AI is, and how big is the share when you think of your Process Intelligence? What does generative AI change, and where can it go?
