Sebastian Steinhaeuser, the COO of SAP
SAP
SAP has two major conferences each year for customers, prospects, and partners. This year’s Sapphire conferences were in Orlando and Madrid. As was the case at Sapphire Orlando, almost all the messaging in Madrid was focused on their approach to AI. While SAP is the largest provider of business software in the world, it is positioning itself not merely as an enterprise software provider but as an AI governance solution provider.
I had an opportunity to interview Sebastian Steinhaeuser, the chief operating officer at SAP, at the Madrid conference. Our conversation centered on AI. The transcript was edited for succinctness and clarity.
Steve Banker: A lot of the conference has been very technical. This is the AI technology. This is how we’re going to manage it. This is the stack. I would like to approach it on a more philosophical level.
As you get older, you start to worry about what kind of world you’re leaving your children. And to be honest, my biggest fear is AI. And it’s not Terminator. It is jobs. I can see AI easily wiping out 20% of jobs worldwide.
I follow the supply chain. We have autonomous trucks. Elon Musk is trying to build an autonomous humanoid robot capable of doing everything a person can in a warehouse or factory. I know there’s been the argument that when one set of jobs is eliminated, other jobs have always come along. But we’ve never seen an automation revolution in which 20% of jobs could be wiped out. Maybe I’m being too pessimistic. That is something that doesn’t seem to get addressed at any of these software conferences.
Sebastian Steinhaeuser: I think about this a lot. I am responsible for our team as we journey to an autonomous enterprise at scale.
My view is that the short-term impact of technology-driven labor disruptions is overestimated, while the long-term impact is underestimated or completely misjudged. Technology always takes time to be adopted. I think this will be a much faster adoption than earlier cycles – every cycle gets faster and faster.
In the long term, well, there is an example I really love. Two of my good friends are radiologists. They finished their studies in the mid-2000s. For the first time in that field, imaging AI became available. This was one of the first real AI use cases – AI for image recognition. It was clear that it solved a problem. It was just a matter of time before AI would take over the majority of diagnoses. My two friends were super worried about going into that field. In Germany at that time, you had to invest money to buy the expensive equipment. And they asked, ” Should we do it?” Maybe there will be no radiologists.
And the prediction was absolutely right. AI now handles, I think, 95% or more of all diagnoses. But we have more radiologists, not less!
What happened? We had completely underestimated the demand for scans. The number of scans just went through the roof. I’m not saying that’s going to be the case for all jobs. It’s just an example of how hard short-term, midterm, and long-term predictions are. But for them, in their 20-year careers, they are happy.
It’s very hard to judge. I think some of the fear-mongering is overblown.
One field heavily impacted by AI is computer science. We still need computer scientists. You still need an architectural understanding of software. It’s not just about being able to vibe code a jogging app. Businesses need to understand how to deploy an enterprise-grade, secure, mission-critical system and how to build and architect it.
But of course, the developer’s job will completely change. Some parts of the job will completely vanish. Other parts will become much more important: creative thinking, understanding the customer, and how to translate customer requirements into a spec. But beyond this, everything from there will be largely handled by AI.
Architectural thinking will become more important. UX (the user interface) may no longer matter because AI generates everything on the fly, as we showed here at the conference.
But does that mean there will be no developers anymore? Would I tell my kids, “Don’t study computer science.” I don’t think so. Because I believe the amount of code that can still be written is infinite. The number of things we can still do with code is infinite.
So it’s very hard for me to judge. At SAP, as long as we are successfully selling more software and providing more solutions to solve more problems, then AI coding is a gift, not a curse, for our developers.
But then, very long term, it’s just very hard to judge.
And of course, there are other famous examples, such as telephone switchboard operators. I think the peak employment for these operators was 10 years after phones that did not require a switchboard were invented. But now, the job no longer exists.
I’m an optimist. I would never bet against people finding a purpose in life to do something that’s rewarding and useful; now, and 50 years from now.
However, I would expect there will be sectors and job types where there will be huge turbulence. Take the industrial revolution; if you worked in a weaving factory somewhere in the US, that job went away pretty rapidly. It also led to disruptions in the economy. But overall, the economy blossomed in the mid to long term.
I don’t know. But it’s a fascinating topic.
Banker: The best thing that could happen is if the transition happened gradually. There was a time when 90% of jobs in the United States were in farming. Now it’s less than 5%. But that transition took over 100 years.
Steinhaeuser: I agree. I think what’s fundamentally different is that the pace is accelerating.
Take SAP. For 50 years, we’ve been driving automation in finance. That was the initial domain that we started in. 50 years ago. Before SAP, people were doing finance and accounting with pencil and paper or punch cards.
Now, with AI, where will we go over the next five years? There will still be finance departments. It’s not that finance as a job will be eradicated. The job will become less and less transactional and more and more strategic in problem-solving, forecasting, analyzing problems, and identifying opportunities to optimize. It’s hard to predict what will really happen.
Take the field of law or consulting. I use an AI tool to challenge me, coach me, and bounce around ideas. So, you could say, “AI can do fantastic legal analysis, I don’t need lawyers and consultants anymore.”
I used to be a consultant. When the internet came along, a radical technology shift, a mentor of mine told me, “Sebastian, we don’t need associates in consulting firms anymore because their main job was to go to a library and do research.” We now had all of this in Wikipedia.
What happened? Consulting exploded.
The same could be true for law. Because of human ingenuity, there will be new problems to solve, new laws to be written.
There do need to be guardrails. There are clearly risks to certain sectors. Security is a risk. But AI can generate significant value and well-being.
Banker: My own job has been impacted by AI. I work for Forbes.com. 50% of our page views have disappeared because readers can search on Google, and the AI says, “This is what you need to know about this topic, these are the main points.”
But that actually has made my job more interesting. I can’t write generic articles. I must write in depth. I have to write stuff that you can’t find on the internet.
Steinhaeuser: What I found for myself, my consumption of high-quality journalism went up, not down. I see a dichotomy between what an LLM prompt generates and high-quality content. You can still tell.
So, that’s what I mean. You can’t say a journalist’s job is now gone; all managers’ jobs are now gone. Maybe we are just massively raising the bar for what we as humans can do.
