This article summarizes an episode of WSJ Podcasts’s video series featuring Christian Klein, CEO of SAP.

Christian Klein, SAP CEO / Photo credit: LinkedIn

Forget building the next big chatbot. SAP CEO Christian Klein believes Europe’s real path to AI leadership lies in its factories and boardrooms. By combining AI with deep industry data that Silicon Valley doesn’t have, Europe can lead the world in “Business AI” rather than playing catch-up in tech.

The necessity of self-disruption

Even successful companies can fail. Klein saw this risk coming for SAP and decided the only way to survive was to make a difficult, company-wide change.

Klein mentions that SAP’s traditional Enterprise Resource Planning (ERP) model wouldn’t last into the future. He knew a change was necessary, saying the company had to disrupt itself because the cloud was clearly the future.

He argues, “You need to disrupt the way you develop a product. You need to disrupt the way you service your customers, deliver the product, and even all the internal functions. It’s a massive transformation, and I was convinced that this is the only way.”

Leading through the backlash
Announcing a plan to change a successful business is a big risk. The move caused the market to panic, forcing Klein to handle the problems with investors and employees.

“In these times, you need to give people confidence,” Klein notes. “It doesn’t help to put even more pressure into the system. You need to be positive, you need to show a clear plan, and you need to overcommunicate. Literally, I was always there; every month, we reported progress.”

How SAP uses business knowledge to stay ahead

That support from his team is the foundation of his AI plan. He disagrees with the common story that new companies will always beat older ones, arguing that in the business software world, understanding the business is what matters.

Klein explains, “All of the startups you are talking to can build an agent on top of such a system, but this agent is really lacking the context. It’s lacking the semantic data context… There’s only one tech company that runs these mission-critical processes end-to-end, and that’s SAP.”

SAP’s advantage is more than its large amount of data
“To understand the business process, you need the logic, you need the data model,” Klein says. “You need to also understand who is allowed to do what. Not everyone in the company is allowed to run every transaction or to see every number, and that business logic sits within SAP.”

The firewall for business AI

This special business knowledge only works if a company’s private data is kept safe. Klein says SAP acts as a needed link between public AI systems and private, important information.

He argues, “These LLMs have no access to a company’s financial data, HR data, or payroll data. And that is what SAP brings to it, so that we can contextualize [the models].”

An LLM needs business data, but providing it is a security risk
Klein says, “In order to understand your question, you need to be able to correlate the chat with the business data, with your supplier data, with your financial data. And this is where our AI foundation kicks in, and that is exactly why we have such a strong belief that SAP will be a leader in business AI.”

Europe’s innovation disadvantage

Drawing on his experience leading Germany’s SAP, Klein believes the lessons he learned at his company apply to all of Europe. He believes Europe’s tech industry is held back by its culture and rules.

EU’s mix of local and central rules makes it hard for tech firms to expand. He noted that the region isn’t actually unified, as companies have to follow a double layer of laws that slows them down.

He observes, “Europe is definitely a place in the world where I feel, oftentimes, we think about the risks first and then regulate before we even start innovating. And I see in the United States, but also in China, it’s going the other way: Let’s first innovate.”

How Europe can use AI to stay ahead
Rather than following other countries, Klein believes Europe should lead the world in using AI for industry. He noted that because Europe is already great at making things like cars, it has an advantage it should lean into.

“We have expertise and data,” Klein concludes. “So we should be the world’s leader in applying AI in order to be the best at running these utilities in the future, in order to produce the best EV cars… Let’s build these models by industry, applying them in our business context, because this is where Europe can still win.”

⚖️ The other side:

• Christian Klein views SAP’s vault of business records as a moat, yet a Harvard Business School paper notes that having the most data doesn’t always guarantee a win. Often, smaller competitors with smarter software or easier-to-use tools can pull ahead regardless of how much information the giant has stored.
• Although Christian Klein pushes Europe to stop building big AI models and stick to industrial tools, studies from the European Parliament suggest that this retreat could leave the continent permanently dependent on foreign tech for the essential brains of future systems.
• Christian Klein believes that startups cannot win because they do not understand the special language of business records, yet MIT researchers found that computer-generated data is allowing new players to train high-quality models quickly and cheaply without needing access to an old company’s private files.

This Too Long; Didn’t Listen (TL;DL) summary was AI-generated and human-reviewed. Read all summaries here.