This series is dedicated to the amazing women in AI, who are not just part of the industry but are actively shaping its future. Join me in exploring their stories, celebrating their achievements, and drawing inspiration from how they're making AI more inclusive and innovative. This week: Helene Staader, Chief Product Owner at SAP AI. Her story underlines the importance of perseverance, adaptability, and the need for more inclusive leadership in tech, reminding us that diverse perspectives are not just beneficial but essential for innovation.

‍

Helene, could you describe a typical day as the Chief Product Owner at SAP AI? How do you balance the technical, product management, and leadership aspects of your role?

A typical day starts with a coffee. I start relatively early, around 6 a.m., so that I can use the first few hours to work through my to-do list and emails. These early hours are also handy for coordinating with one of our Asia teams (we have teams in Singapore and India). My meeting marathon starts between 8 and 9 a.m. From then on there is really no discernible pattern, except that I spend a lot of time in meetings. A lot needs to be coordinated. Especially now that AI has become such a focus at SAP, as evidenced by the latest reorganization at the beginning of the year. Our newly founded SAP Business AI department has ambitious goals and needs to be restructured in order to be solidly positioned to meet the new challenges. We talk in various rounds about the strategy, about the technical implementation, change our processes in order to be able to scale better, and redistribute the topics among the teams. That is the challenge, but also the appeal of my job as CPO. It is so versatile and very varied.

First and foremost, I am responsible for shaping the vision and strategy of our products together with our PMs and implementing this vision with the development teams. This means maintaining contact with our stakeholders, prioritizing their requirements, initiating the technical implementation of these requirements, then accompanying the development process and ensuring that our products comply with SAP standards. There are many processes to consider and follow. And of course I don't do any of this alone. It's teamwork with our PMs, architects, POs, program managers and people managers. And that’s incredibly fun. I'm not always and everywhere deeply involved in the discussions, but ultimately everything comes together for me. And although I don't have any direct HR responsibility, CPO is an absolute leadership role! Good processes and a clear strategy make an important contribution to satisfaction in the teams. And I am very aware of this responsibility. My role therefore combines these three important aspects: product management, engineering and leadership. For me it is an absolute dream job.

You mentioned before that your journey into AI wasn't a direct one. What challenges did you face transitioning from a developer to a role more focused on AI and product management, and how did you overcome them?

It was constant growth, growing into new topics and new roles. 20 years ago I started my career as a developer. During this time, I gained experience as a Scrum Master and as a project lead, which deepened my knowledge in the areas of software engineering, agile software development and coordination. I got my start in the AI world in 2015 when I was given responsibility for a small AI component. A short time later, I applied for a fellowship in SAP's newly founded central AI team, which aimed to build an AI platform to make it easier for developers to get started. I quickly realized that this was exactly what I was interested in. We do not build AI methods or applications ourselves, but it is essential to understand how data scientists and AI engineers work, what the current trends are, and the existing obstacles in the development of AI applications. After my fellowship, I joined the team as a product owner and have stayed with it to this day. Our team grew from around 10 people initially to around 150 people, and at some point I took on the role of CPO for the entire division. Each phase of this transition brought its own challenges: How do you manage such rapid growth? How do you deal with high visibility and pressure? How do you stay current with technology in a rapidly evolving field? And how do you optimize your time management in the face of increasing workload? Sometimes training helps, other times coaching, but above all a lot of self-reflection and the will to develop further. Another important factor in overcoming all these challenges is management support. I was very lucky to have managers who gave me advice and support.

In your view, what are the most important emerging trends in AI right now? How do you see these trends evolving in the future?

Of course I'm not saying anything new, but for me, generative AI is the greatest technological achievement of this century - and not just in the area of artificial intelligence, but in the entire technology landscape (so far). It's a real revolution! So what’s the big difference from previous AI technologies? From my point of view, the difference lies in the simplicity of use and the broad coverage of use cases with just one model. Generative AI models can be used out-of-the-box, or you can achieve great results with relatively little effort - through prompt engineering and grounding. Data scientists are no longer needed to develop such cases, which makes entry into the fascinating world of AI much easier for all companies. I believe that GenAI will also have an immense impact on the way we work in the future and how we communicate with computer systems. Obtaining relevant information will be even easier, and routines such as composing emails or summarizing meetings will soon be fully automated by AI. And that's where we come across the biggest problem with generative AI: Can we trust the information and content generated? For example, is it possible to publish a generated job advertisement without reviewing it? The GenAI models are prone to hallucinations and are not free from bias. So our biggest task will be to handle the generated output carefully. We have to enrich the models with domain-specific knowledge (grounding), correct known biases and, above all, always indicate when it is generated output. The “Human in the Loop” principle is still absolutely important and indispensable at the moment.
It's very exciting, but also a little scary, how quickly AI is advancing. However, our ethics guidelines for AI and the associated regulations are lagging behind. But honestly, I don't think taking a "time out" to think about how we should deal with it is a realistic solution. So we have no choice but to keep these aspects in mind when developing applications with AI and course-correct whenever necessary.

Based on your experiences, what do you think companies need to do more of to support women in building successful careers in the tech industry?

Companies are practically at the end of the chain of institutions that influence women's self-image and their character. The influence begins in the family, continues in kindergarten, school, training center, etc. Unfortunately, women in our society are often denied technical understanding. And if you grow up with such views in your immediate surroundings, you unfortunately often adopt them without being asked. In addition, certain tasks, such as childcare and caring for relatives, are mainly assigned to women. As a result, not many women find their way into the technology industry. It is therefore important to enable women to develop freely and without prejudice right from the start! In order to achieve this, there needs to be a rethink in society as a whole, more diverse role models for girls and thereby the removal of mental barriers in relation to their inclinations and abilities so that they can choose their own path in life (with or without children). Two very important ingredients in our society are necessary for this: tolerance and acceptance. It should be a real choice and not just a theoretical one. The good news is that a lot has happened in the last few years and the rethinking has begun. My children's generation is already much more open than my generation was, and I hope this trend continues.

Now to what else companies can do: Good coaching concepts and training can have a huge impact. At the beginning of my professional life, I took part in a 'Gender Awareness' course, which offered many 'aha experiences' and in which I reflected on myself for the first time in comparison to my male colleagues. This does not mean that women have to adapt to a 'male world' - on the contrary: women often complement themselves with their characteristics, which in my opinion contributes to a healthier working atmosphere. But when you become aware of certain mechanisms, you can control certain things better (increase visibility, present your successes, etc.). Managers also need to be trained to help women succeed. And then there are very banal things that are still not a given: enabling childcare, offering flexible part-time work (if possible without compromising your career).
From everything I have said, I assume that discrimination is already a thing of the past. If this is not the case, that is the first thing every company should change! I would now like to extend this statement to other disadvantaged groups. There should be no place for any kind of discrimination in a modern company - regardless of whether it is based on gender, sexual orientation, skin color, religion or disability. The world is colorful, and that should be reflected in every company!

You emphasize the importance of diversity in teams. How do you think diverse teams contribute differently, particularly in AI and tech fields?

Oh yes, I'm a big advocate of diversity in teams. On the one hand, I am convinced that this will enable us to develop better software. We may build software for companies, but ultimately we build it for the people who work with it - and those people are diverse. The more diverse the influences that flow into development, the more likely the end product will meet the needs of many and not just those of a small group. On the other hand, it's just a lot of fun to work in such a team! You're constantly learning something new: new facts about different cultures, different mentalities and perspectives, different approaches. You expand your horizons every day and change your world view. I am convinced that we all become more tolerant people when we work in diverse teams. And of course this diversity not only applies to different cultures and genders, but also to all other groups in our society.

In your experience, what are the most significant challenges organizations face when adopting AI to transform their operations? How do you think should companies approach these challenges, and what strategies can be effective in facilitating this transformation?

In fact, my experience so far is that, despite the great hype surrounding AI, companies are often hesitant to use it on a large scale. In my opinion, several factors play a role here. First, there is the difficult assessment of business value. What specific benefits do I have as an entrepreneur if I automate certain processes with AI? It is clear that AI can save manpower, but since it only provides the 'most likely' solution, there is always a chance of error. So the question is: What error rate should be expected and how difficult is it to correct these errors? These questions are highly specific and the answers vary depending on the use case, model used and integration into company processes. Another problem is simply: trust. How did the model arrive at its solution? In most cases, the model unfortunately cannot provide an explanation that we can understand. Although there are approaches to improving this explainability, these have not yet been possible across the board and satisfactorily. The last point I would like to address is the feeling of being overwhelmed. Many are overwhelmed by the number of different models and tools (this number is growing every day!), their quality, legal requirements, data protection and ethical guidelines for AI. Very few companies have enough of their own AI specialists to overcome these challenges. And that is exactly the point at which we as SAP help our customers. We integrate AI into our products and use selected technologies that we have reviewed and found to be good. We would also like to offer this selection to our customers who want to build their own applications and provide them with decision-making assistance depending on the use case. We also want to enable them to integrate these new AI applications into their processes and our products as easily as possible. So much for challenges. But what are the solution strategies?

‍
The first step for a company to advance the use of AI would be (as banal as it may sound) to identify lucrative use cases. However, this is by no means as trivial as it sounds. What brings the biggest improvement or savings? The next step is to select the right technology, and this depends largely on the company in question. Do you have the necessary skills internally to make such a decision? And if not, it is often advisable to rely on the support of a competent and trustworthy partner.