Do not underestimate how important humans are, even as we automate, the chief technology officer of SAP says

• Juergen Mueller, SAP's chief technology officer, told Insider how SAP was helping clients innovate.
• SAP's clients are focused on cutting "time to impact" and responsibly relying on AI and automation.

Juergen Mueller is the chief technology officer at the German software giant SAP. The following has been edited for length and clarity.

For a company that creates software to help companies run their businesses, there's never been more to consider. Automation, generative artificial intelligence like ChatGPT and Bard, and machine learning are pushing business leaders like never before.

Juergen Mueller has to think about many of these issues. Mueller, SAP's chief technology officer, explained to Insider how the German company — which focuses on cloud-based subscription services for its financial-reporting, inventory-tracking, and human-resources applications — was balancing the pressure to drive innovation with the need to make progress responsibly.

What keeps you up at night as you push both your company and your clients forward to meet the demands of the new world like we're living in? It's so different now than even a year ago, and we find ourselves at the third anniversary of pandemic shutdowns hitting the US.

The answer is two things: Companies are pressured a lot and leaders are pressured to deliver very quickly. And we also got used to that because in the pandemic, actually, we could change very quickly, for example, to remote work or hybrid work. Within a day, you switched, and then within days or weeks, you needed solutions.

One that actually quite a few of our customers had was when they were hiring people into remote work. We want to give them a kind of allowance when they start for not just the normal equipment they have but also an external monitor, if they need a desk and chair, drawers to take care of work safely and ergonomics — and then the fastest customer I got to know implemented this within two weeks.

Time to impact — this is super critical.

Then the second thing I observed is user experience. We're mixing, especially now, the home and work experiences and, of course, I want my business system to behave like my iPhone, iPads, Android, whatever you use, because it's pretty slick and it works flawlessly. That is also how I want my business system to behave.

For example, we had a customer that produces wind turbines, and they created a mobile app for their 10,000 field-service technicians. So they have like a million wind turbines installed, and regularly, they need to maintain them and check them. And this is the No. 1 requirement: user experience. They deployed this app, built on SAP Business Technology Platform, and said it saves 400,000 hours per year. Wow. Their workers are now more efficient.

Really, we need to push, we need to understand users and "what is the job to be done," and then really support them with information and data.

We're only in March, and AI and layoffs have dominated the business conversation this year. Can you talk about how you're using AI to inform decisions, create innovation, yet also not scare people? With the digital transformation that's going on now, the question is, "How fast can we do it in a responsible way?"

We are really embedding AI in our products. We have, like, 130-plus AI use cases, such as the finance area. Or in retail, how to identify which prices to pick out and how to identify markdowns or how to identify slow-moving goods that, actually, a company should promote. And we're now talking about promotions marketing. We work with vendors on Cyber Monday or Black Friday and easily have more than a billion personalized messages. What information do I send to you that you're interested in? Maybe it's a bike company: You're interested more in road bikes or fun bikes. Today, this is driven by AI, but then the content is still driven manually.

Generative AI like ChatGPT, of course, will have a major impact, but you still will have humans in the room. The tech will not be defined, not for the next couple of years. You will have humans in the room, but maybe now you can be 10 times more finely granular because you only need to spend 10% of the time for fine-tuning the message, etc. And by that, we will see the more hyperpersonalization or things like supply-chain routing. We have driver shortages, we have container shortages, chip shortages, and product shortages in some areas. How do you substitute one product with another so that you still can produce? These are super-complex problems where we have use cases embedded in our software.

You talked about providing and embedding AI and machine learning and giving customers ways to solve problems in a scalable, ethical way. Can you unpack that a little bit more and talk about what you're seeing customers desire or their needs? And how are they solving these problems in a scalable, ethical way?

Imagine some simple things but still very relevant things. We work with the largest retailer in Australia, and they have 4,000 stores, 40,000 items per store, and they want to have a one-year view into the future — which items they are likely to sell in which store. They want to drive their supply chain to have these goods in those stores available to reduce waste, for example. They use AI and forecasting mechanisms in order to do that.

But you can imagine if you have such a plan and then a pandemic hits, the plan doesn't work anymore. So we also provide, together with partners, tools to constantly monitor how well do your machine models still work because it also changes.

We were, I think, even the first European company that introduced an AI-ethics board. It's a board composed of internal and external people. We created guidelines. We train all our developers who have to do with AI, for example, to make sure we don't have bias in training data. And we also have clear rules about what we will do and what we will not do.

We would, for example, not develop a system that would be firing people. You could, in theory, use AI to create a model to recommend whom to fire. We have standards in our understanding of how our technology should be used.

Automation is something that all companies are doing to save time and money. And there are a lot of benefits to it. On the flip side, workers are worried about their jobs. How are you and SAP helping your clients drive automation within industries without costing jobs and making sure the right message gets across on the benefits of automation and how that can lead to upskilling if the company does it right?

We have companies automate their end-to-end business processes. Of course, with all technology shifts, as you mentioned, there's also downsides. Some capabilities will be obsolete, and new capabilities are created, like machine learning. When machine learning arrived, first that automates a lot of capabilities, but now you have completely new jobs.

How do we help and strive to help companies automate and streamline, but not underestimate human ingenuity that is necessary in many, many different areas? And this will also not go away that fast, even though we saw layoffs.

In many, many areas, there's a shortage of labor. There's the desire to work less and have at least a little bit of a work-life balance. So therefore, we have to also have an eye on these issues, and, in particular, if you automate one area, what I think is most important — and this is also what we tell our clients all the time — educate people. Help them level up their skills. So this is a huge topic. How can we provide continuous learning to employees?

Do not underestimate how important humans are, even though you automate.