top of page

Artificial Intelligence: Turning hype into outcome

MindSpeaking Podcast Episode 8 - Andreas Welsch, VP at SAP



๐ŸŽ™๏ธ Listen to the full podcast:



Listen to the MindSpeaking Podcast:

Spotify

Apple Podcasts

Google Podcasts

Other Platforms



Connect & free resources

๐Ÿ‘จ๐Ÿปโ€๐Ÿ’ป Gilbertโ€™s Linkedin

๐Ÿ”‘ Free email course

๐Ÿ“ˆ Discover your Data maturity level

๐Ÿ“™ People Skills for Analytical Thinkers




Introducing Guest



Gilbert Eijkelenboom:

Today on the mind speaking podcast, I talk with Andreas Welsh. He is a VP at SAP and he has more than 20 years of experience in the software. Industry. His expertise is an artificial intelligence and Process automation. And he's turning hype into outcome is sharing his insights with his intelligence briefing via LinkedIn and his newsletter. And what we're talking about today is how to make AI projects a success, what type of mistakes to avoid, and how to foster AI mindset. Also, we talked about the role of empathy, communication, and a lot of tips how you can do this in your own organization. So I hope you enjoy this episode with Andreas Welsh. And Andreas, nice to see you again.


Andreas Welsch:

Hey, Gilbert, thanks for having me.


Gilbert Eijkelenboom:

I'm really looking forward to this conversation today. And we're gonna dive into a lot of different topics. For the people that are watching. They can see your background and I've not had many guests that are that have such a fantastic background. So I'm happy to see that in the first place.


Andreas Welsch:

Thank you so much. I'm really passionate about getting the topic of AI across them. So I wanted to make it as comfortable and as cozy as possible. Yeah, show a bit of personality.





Who is Andreas Welsch?



Gilbert Eijkelenboom:

Yeah, fantastic. I think it's going to be cozy conversation about AI SE and AI, light blinking in the background or clear in the background, and I'm looking forward and of course, we're gonna talk about AI but I'm also curious about you as a person. So I would like to hear a bit more about you. Where did you grew up? How was how was it for you to grow up? And yeah, tell us a bit more about yourself? Sure, absolutely.


Andreas Welsch:

So as you can probably tell by my accent, although I live in the US, that's not where I am originally from. So I was born and raised in Germany, on the western side of Germany. Very close to the border. Fence, actually only about two three kilometers away. In always lived in kind of a region also lives in the south of Germany, continents, which is right on the border with Austria and Switzerland. Lots of beautiful scenery in the Alps and you know, always different different types of cultures are living on the east coast in the US and Philadelphia, would still consider that a bit of a border region with Pennsylvania, Delaware, New Jersey being close by certainly not as, you know, the same kind of different countries around but certainly you'll also get a different different perspective and a different flavor. And so, you know, growing up, I had a couple of different things I wanted to become. At one point I wanted to become a pediatrician, and a bunch of other things, but it wasn't until I did my first internship in a company in a software company actually that I got exposed to software it and all of that world. So that was really like in my late teens. Whereas what what you can do with it with software, what impact you can have and that's really set me up on that trajectory and was just the starting point for me to learn more about all of that.





Becoming a People Manager



Gilbert Eijkelenboom:

He also mentioned that to become a people manager, you need more than technical skills. Can you share more about that? Oh, but your thoughts


Andreas Welsch:

you know, I spent a good 14/15 years in it in my career. That's that's where I would say I grew up personally and professionally. So I had the fortune to work with many great people and uncontained and work with many great leaders and work for many great leaders. But when I saw that I wanted to take on more responsibility. I was kind of hitting a glass ceiling at the time, nobody else above me was moving up or was moving out limited opportunities. And so I said to myself, Hey, by the time I'm 30 I do want to either be a people manager in it, because that's really my was my main subject matter expertise was at that time, or look for other opportunities within the same company. And so that was that was what I ended up doing because there were no opportunities in a short amount of time or a short period of time that I could presume. So I you know, I already knew about data centers and servers and networking in some some bits of automation and scripting. So it was a good way to take that as, as as the foundation and build on it and had a fantastic opportunity and SAP. Were right currently still work to build on that to do what we call a fellowship, which is basically like an internship for a couple of months where you can join a different department learn about what they do, but bring your knowledge to the table as well. And then at that time, expose me to the cloud. And it's like the early 2010 summit 2000 Tim's what customers were thinking about how you build this cloud thing, and again, from there, they led me towards CTO. And at the time 2015 16 digital transformation was just starting. What is it? How do you talk about it? Why does it matter? Again, how does cloud fit into that picture? And so, the next step after that has been similar. So for me, the approach has always been take something that you already have an expertise and that you're pretty good at and see what other new topic you can add to it and where you can learn and where you can grow. So go outside of your comfort zone, but I would say don't delete it completely. Don't leave it behind. You know, build on your strength. Keep learning in keep looking for something





Going outside of comfort zone & building new types of skills


Gilbert Eijkelenboom:

interesting. So what you're saying is that you need to build on what you already know, right? Your current experience, so maybe one foot within the comfort zone but also wants certainly one big step outside of it, and how do you think, I guess many people would agree, I think with that sentiment, how do you think companies can adopt such a culture of going outside of your comfort zone and, and building these new types of skills?


Andreas Welsch:

I think what's really key is creating the learning culture giving those opportunities, and certainly not every company will have such a program where you're encouraged as an employee to move around for a couple of months, you know, in your career, but it can be as simple as as a virtual community. Where is your back in the office meet, have coffee with other like minded folks or just exchange? What is it that you do and maybe what is sales look like? Or what is pre sales look like? And what does a day in your life look like and what my life looks like so just get an overall better understanding and see if there are opportunities to collaborate or if not at a minimum learn about somebody else's area and expertise and see if that interests you or if it's something that you want to enter into more





Important traits or skills of a People manager working in a tech area



Gilbert Eijkelenboom:

and, and then becoming a people manager you mentioned, what what do you think are some important traits or skills you need to learn as a people manager working in a technology area?


Andreas Welsch:

So in one of my last roles, I live what I think can be best described as an AI Center of Excellence. Now, I'm not a data scientist. I'm probably not even a data person, or don't have a background in data. But I've done a lot of work on automation in this kind of thinking is something that really excites me. So I think having a good understanding of the sub subject is obviously key. So you know, what you're talking about and what to look out for and as you build your team, what types of skills to look for, but I think as a as a people manager, and it gets even more so, you know in pandemic post pandemic world, whatever we want to call it, and it's normal, the new normal. A lot of us are still working remotely where we where we have that opportunity working in, in technology. So empathy, I think is a is a key one because, you know, we were juggling not only our work life, but certainly our personal life as well at the same time. So, not everybody might be available at the same time. People have different work schedules. Now, especially with different things on their mind, whether it's not having been able to see relatives or loved ones or friends for a long time, not being able to go harder, you know, even health topics that have become more or come more to the forefront. So overall that empathy that understanding I think is key, but also setting a clear vision of what you want to achieve. In its I'd say it's fairly easy if you're an individual contributor to just join along for the ride, float along it, somebody else is coming up with the ideas and we just, we just do what we're told. Which is one mindset right? As a leader, certainly you do want to set that vision. So what is it? Where are we that we want to go to then do we want to move towards what are some of the big trends that we're seeing? What's the impact we are making that we can make, and how do we set up our teams? What are the strengths that we have where do we have some some gaps and how can we fill them whether it's developing our people in that area, or bringing on additional ones that can complement that skill set, I think, between empathy, vision and certainly also execution that's a good trend.





Buzzwords



Gilbert Eijkelenboom:

Yeah. I agree. It's a great triangle and to create that vision and to communicate that vision, you should use different words than buzzwords right because the link that I want to make, as you mentioned, a lot of times as you're looking beyond the buzzwords what does it mean to you?


Andreas Welsch:

So look like buzzwords are not there to to create excitement. It's like this hype right? Everything is AI these days. But when you pull back the curtain, is it really or is it machine learning? Is it predictive? Is it statistics is a rules? So I think being being more realistic, and more open and honest about it, and creating that transparency, what it actually is, is really important in doing that in simple language, so that you understand it you can communicate it but more importantly also the person that you're talking about, sorry that you're talking to, can understand it in a lot of times in data roles. You are the technical expert, when you're talking to somebody who's probably not that familiar with terminology with these buzzwords. So what does aI mean to somebody in finance? What can it do? Why should they care about it? Well, hey, instead of you having to process incoming payments every day and look through a list of hundreds of items, say Well, I think this one matches that invoice. And this one matches it too. And that's a big part of your job today. If we can bring in technology to help you with that, and all you need to look at are the five or 10 instead of the 50 or 100 things that come in if you only look at the exceptions, I think that changes the trajectory for somebody in a specialized role. But giving an example like that makes it also a lot more tangible. Than you know, the misconceptions that exist around Terminator Minority Report The Matrix in everything that the refund seeing from Hollywood, like it's not that it's a lot more concrete, and it can be a lot more impactful.





Cultural Changes



Gilbert Eijkelenboom:

Yeah. And I like the example that you gave because it shows how if we make it very concrete, we can bridge that gap between it or data and business folks who don't, who may not have the technical expertise that people in technical roles do. But I think that's how you can communicate and bridge the gap and get closer to the to the other side. Absolutely. But I'm also wondering about because you're you're talking a lot on LinkedIn about the cultural training changes that we need to make and AI projects, you know, how to make sure they're succeeding and that we need to treat them differently compared to IT projects. What are some of those cultural changes you think we need to make in companies


Andreas Welsch:

I think the most important one is creating an awareness then AI is so much more than just technology, technology and data have a very, very strong core and component and our center to it. But there's so much more around it. Whether it is skills and skill sets to help again the right people in the right skills. If not, what can we teach them? What can they learn? Where do they find information? Its governance in data. What type of data do I have where I get it? Where do I get it from, but what am I allowed to do with it? Is a personal identifiable information in it. How do I annotate data? How do I create better training data? What do I do if I don't have enough data? Lots of things around that topic. If you if you build this for your own company internally, they're obviously work with stakeholders in different departments and domains that might not even be as familiar with AI and data and other things that don't matter here. So you also become a bit of a multiplier. I think you'd need to create a bit of awareness or that mindset and AI literacy in the company that will help you build projects in a better way and hopefully a more successful way in the thing that I've seen work very successfully is if that happens in close collaboration that happens in Canada because a lot of times if you are working on data and data science, your technical experts again, that's perfectly fine. It's perfectly good. We need that deep technical expertise, but not everybody can be an expert in everything. So we still find that a mix of deep data science expertise, and for example, deep finance expertise or deep manufacturing expertise remains to be rather rare. So when a pair of the two skill sets up, somebody that's an expert in manufacturing that knows how the business process works in your company, and somebody that's really good and data and has deep expertise and put them together so it's almost like a greenhouse effect, but they both can grow together.





Collaboration


Gilbert Eijkelenboom:

Yeah, that sounds like a great approach and how, how do you think that could work in practice? I mean, I see that it's very effective to combine those two people that have different experiences and different skill sets. What I also foresee or what I have seen is that when two people work together with different expertise, or maybe even different personality, it's more difficult to create a very strong collaboration. What have what is your experience and how do you think we can do this? Well,


Andreas Welsch:

what I've seen been very important is sponsorship sponsorship from the top, then, hey, this is a priority. We need to do this. We want to do this. And we need different people to come together on their team, and it won't be perfect. It won't be perfect the first time and probably not the second or third time. But it's a learning opportunity. So there's something to gain for both in obviously for the company as well that asks you to run this as an employee. So with that mindset, it's a it's a growth opportunity. And you share a little bit and I share a little bit and at the end of the day we're both smarter about what you know what your domain does and how AI and data can apply to it. And the other way around, I think that's a fantastic opportunity. But a lot of times really it needs that sponsorship from the top that this is strategic. This is something that we want to be successful in as a company.





Mistakes that organizations make



Gilbert Eijkelenboom:

But but something I'm also wondering about this, if you talk about AI programs, what are some of the frequent mistakes people or organizations make when setting up these AI programs and, and what makes them more successful? Because you've touched upon the mindset, the collaboration between different types of people that have different experiences and knowledge levels. What are some of the mistakes that you see organization make?


Andreas Welsch:

So there there are a couple of things. Maybe I'll start with the ones with the examples where I've seen this work particularly well. In in those examples. The customers that I have worked with in previous roles really defined what the objective was, from the very beginning. Why are we looking at AI? What do we want to achieve? In many of these cases, it was around shared service centers in the finance example that I gave, for example. So naturally, there is concern, there's maybe a bit of distrust, if you put in this automation at this AI thing, doesn't mean I'll be out of a job in a couple of months because that thing you tell me can do the job much better and much faster and much cheaper than I can. So setting the expectation that hey, that's actually not what we want to accomplish. Yes, we do want more automation. We do want to reduce our costs. But we also need you and we need your unique expertise, not only to look at the exceptions, but what would you do if they 60% or 80% of your time was now freed up because we have this AI thing that takes care of the boring, mundane stuff. And you can follow up with the customers insane. So you were supposed to make that payment a week ago, you still haven't documented what's going on in your in your business are you alright should we change the change the terms when can we expect the payments so these kinds of things, you know, where you have a much more personal connection and can a human connection, let the I really worry about the other things. So getting that point across that it's not to replace people, but rather to augment them and to help them genuinely, genuinely do that. That's That's key. The other part is, to your question, what should people not do? I would say do not treat AI as any IT project is any other IT project because it is so much more complex, like something that people Dimension Data, Data privacy dimensions, how do we get access to data? Certainly a business dimension as well because we're not just doing that in a vacuum or for for our own passion of seeing how how we can get the last percentage points when we when you tweak that model, certainly all good and relevant aspects but at the end of the day, we do want to drive business outcomes. with it. Again, increase automation, reduce repetitive tasks, maybe get better insights, give better recommendations to your customers, whether it's like the classic examples like Netflix and better recommendations for movies, or in an E commerce scenario. Better recommendations for products are combinations of products, or upsell customers who've bought this item also bought that but really it's about the value that you want to create for the business. When all of that comes together, and maybe even when you want to market that data product externally and when to bring in your marketing and sales teams and how you do that. That's really a 360 degree view. I would say take that approach in make yourself aware that it takes a lot more than just data which is still at the core





Automation



Gilbert Eijkelenboom:

right. What I noticed is that you talk about empathy quite a lot. So in the beginning about when you become a people manager of course you need to have empathy towards your your colleagues or your the people you work with, because they have a different situation and also empathy towards business stakeholders who are not as technical and also to people that might be might be afraid for getting replaced by the by AI. So I think it's it's such a important concept and understanding the perspective of other people. And I also like that you've pointed out that automation and AI is not only arise in, you know, using technology, but actually it can aid the personal connection because you have more time to have that personal connection. And I think is a very important point. Absolutely. If we're talking about automation because you spent a lot of time in automation, you're definitely a fan of it. But you also mentioned that just because you can automate something, doesn't mean you liked the shoot. Can you expand on that and elaborate on that?


Andreas Welsch:

So I see a lot of times when we when we look at technologies when we talk about technologies, we're so focused and narrowly focused on what do we have and where can we use it. But again, it needs a different approach, right? If you take a different approach, chances are you're much more successful because they have a much broader field of view. And range of view of what you can do and what do you want to do. So what I see a lot of folks do is hey, we want to automate. Okay, what are the low hanging fruit? What should we do what's first but second. Getting too quickly. into motion mode in wanting and needing to deliver that return on investment. But I think if you take a step back and you say well, we could actually do this, but does it add value? Does it make somebody's life easier? Do we expect to see a certain return on investment? When do we see that? What maybe business KPIs and metrics can we positively influence and can we measure that? I think that overall to me, is a lot more important in is a lot better. Gives you a lot better chance of success and say, Hey, we cannot amend this. We can automate that. We can automate that. And then you end up with a lot of waste because you really didn't have to automate it. Or there wasn't even a need for this process step tasks to be there in the first place. So think take a little more holistic view. What do you want to accomplish? And then select the right tools for the job. And sometimes, you know, it might not even be AI or it might not even being an automation or robotic process automation. Sometimes it's as simple as a role or organ thing. Do we actually still need that thing? We've been doing that for 20 years? Does anybody still know why we're doing that? And a lot of times, the answer will be Oh, yeah, you know, John or Bob implemented that when we went live with this or that system. I actually don't know if we still need it or why we have it. So that's also a good opportunity. And to to, you know, to clean house and breathing things again. Yeah, I like that.


Gilbert Eijkelenboom:

I once read about doing a reverse pilot. So instead of introducing something to see if it works, to delete something and take something out and see if people start complaining and I think for many automations This could also work and I see myself in my personal life also tried to automate stuff or my calendar or productivity things that I spent a lot of time on, that actually take more time setting up the automation and maintaining it then the benefit I have in terms of productivity. So I even see it in my my own life. And I think it's important enough to always take a step back and understand okay, what are we trying to achieve? And also regularly check if it's still still valuable?


Andreas Welsch:

Correct. That's where I think our excitement sometimes clouds our judgment.





Role of communication skills in bringing AI programs to a success



Gilbert Eijkelenboom:

Exactly, exactly. What I'm curious about this because we spoke about AI projects and what makes them successful or what are some mistakes, what is the the role of communication skills in bringing these programs to a success? And I think I think we see these several levels right sponsorship down to within teams. What do you think are some of the detail changes that we see or what is important when it comes to communication?


Andreas Welsch:

So I think the the part of the important is that is adequate to the stakeholders that you're trying to address that you want to share information with, certainly at the top management senior management level, it's about overall high level concepts, business impact, that you want to drive. What we said about mid level or maybe first leadership, how can we collaborate? How can we bring people together and how can we free up some of the time so that they can collaborate? And what's in it for me, what's in it for you? What's in it for us? And then at the individual level, and it's why is this important? Again, why does it not replace your job and help you actually be more efficient, effective, engaged in the end? I think at a minimum that's, that's that's key. But there's one one story I want to share one example. So last year, we went through a number of pilots internally, things that we wanted to build into our own software products that we sell to customers and as SAP. And so it was an interesting time because it was just just before and between the summer holidays people are coming in, as we're just about to leave and quite frankly, we didn't have a lot of time to prep. But we wanted to meet with our head of architecture in head of portfolio, who at the end of the day assigns money to the projects. So we had a meeting where we brought in our data scientists to explain some of the the intricacies of the projects that they were working on and the progress we were making. The challenge was that we didn't have a lot of time to prepare. So we said okay, between your own vacation coming in you going out. Let's figure out what we do want to show. The challenge we have that in the end, we ran out of money, and we get stuck somewhere in how we optimized the model and with some of the features and parameters where we didn't really get to that point. When one of our executives had to move on to another call. It wasn't really stressful meeting. So in terms of communication in these things, the question is, what is the key information and how quickly can you get that across? And what is the key information for that stakeholder that they need to know to say, yep, that's great. I continue funding this or when roadblocks Do you see where can we help what what else do you need? You know, sometimes things are right. And that's okay. But I would also say take time. And again, think about who's your stakeholder? What's the story you want to tell and what's the outcome you're going to achieve? And rehearse, rehearse and rehearse? Yeah,


Gilbert Eijkelenboom:

it's very rewarding. Part is also important, right? And getting getting learnings from every interaction because every time you learn something and you try something new, see if it works, if it does, or if it doesn't, if you successfully persuade someone, or get someone in the right direction. And I think we learn about those interactions every day. What I don't have to do right now, but I will in the future, hopefully. But I want to hear from many people who do have gets to learn a lot from kids and I know you have two kids. And actually they they kind of help you to create content on LinkedIn, because they wake you up very early, right.