How Do You Move AI From Pilots to Programs? Ivo Strohhammer on Making AI Stick

Andreas Welsch: So everybody talks about AI. Everybody talks about it as a shiny object. We need to do something with AI. But very few actually have a plan for how to go from pilot to production, and not just with one use case or scenario, but with a whole program. My guest today has deep experience doing this for large enterprises, and he’s now helping small and medium-sized businesses do just that. So stick around as we figure out how you can go from pilots to programs and make AI stick.

Andreas Welsch: Welcome back for another episode of What’s the Buzz, where leaders share how they have turned AI hype into business outcomes. Today we’ll talk about how to go from pilots to programs and make AI stick. And who better to talk about it than Ivo Strohhammer, who’s been doing that for a long time. Ivo, thank you so much for joining.

Ivo Strohhammer: Thanks for having me, Andreas. It’s a pleasure.

Andreas Welsch: Fantastic. Ivo, when I published the AI Leadership Handbook in the fall of 2024, you were the first person to comment on LinkedIn and say, “I see so many great examples there. I’ve been voraciously reading it and there are so many parallels to what I’ve seen.” A really good connection and friendship has evolved over the last year and a half, roughly. So I’m super excited to have you on and to learn more about what you have seen. But maybe for those of our listeners and viewers who are not quite familiar with you yet, share a little bit about who you are and what you do.

Ivo Strohhammer: Of course, thanks. My name is Ivo, as you can hear. Swiss native, living in Switzerland, working for Siemens, driving the AI adoption there. We started two and a half years ago with programs, and we were desperately looking for some help or guidance because it was uncharted territory for everyone. Then your book came around, and that was really a helper for us to make the next steps. Meanwhile, after doing this for years, I decided to move on to one of the next steps, because let’s face it, the big corporations have addressed it, they have solved it to a certain degree. The next step is really to bring that to the smaller enterprises, to the community. So we are building an ecosystem here in Zug, locally in Zug for Zug, to really drive this adoption, to make it more from hype to value at the end of the day.

Andreas Welsch: I love it. I think it’s so important. I’m seeing similar things here in the US as well. When I visited you in Zug, just south of Zurich, in the summer, it was amazing to see the vibrant ecosystem that already exists there across so many different industries. Financial services is certainly one of the prominent ones, but many others as well. So I’m excited to see what you’re building up there, and I know it’s a very supportive area and community. Folks, if you’re just joining the stream, put a comment in the chat where you’re joining us from. I’m always curious to see how global our audience is. And also, don’t forget to pick up my new book, The Human Agentic AI Edge. It’s out on Amazon now. It’s all about how you can shape the next generation of AI-ready teams when AI is entering the workplace, when we’re seeing more AI slop created by people who don’t really know how to use AI well. You likely see this in your inbox all the time, but as a leader, how do you empower and encourage your teams to use AI and use it well? You’ll learn the tips and tricks from more than 50 interviews with AI leaders in The Human Agentic AI Edge. So make sure to check that out. Now Ivo, in good old-fashioned What’s the Buzz style, should we play a little game to kick things off?

Ivo Strohhammer: Wonderful.

Andreas Welsch: So let’s do this. When I hit the buzzer, the wheels will start spinning, and when they stop you see a sentence. I’d love for you to answer with the first thing that comes to mind and why, in your own words. You have 60 seconds for your answer to make it a little more interesting. So, are you ready for What’s the Buzz?

Ivo Strohhammer: Yes.

Andreas Welsch: So, here we go. If AI were a color, what would it be and why? 60 seconds on the clock. Here we go.

Ivo Strohhammer: It’s mean when everything is colored blue in the background, right? The first thing that comes to my mind is that meme where now with AI everything is a unicorn and rainbowy and shiny and colorful. So I’m thinking about rainbows at the moment, multicolored, where everything now becomes more colorful. AI can be everything. At the moment it helps us in bots, on a screen, in wearables, in machines, everything, right? So it’s…

Andreas Welsch: As colors rather than a color. I like that. And sticking with the rainbow theme, ideally at the end of the rainbow there’s a big pot of gold that everybody is hoping for. But let’s see. I haven’t found the pot of gold yet, at least not at the end of the rainbow.

Ivo Strohhammer: Has someone…

Andreas Welsch: Yes. Right. But you know, sometimes that’s maybe also what organizations are chasing here, the proverbial rainbow. So, all right. Awesome. I love your answer. We stayed well within time and it was definitely very colorful, but you still need to know what you’re actually building for it to have value. Looking at the last two years in AI, last year we obviously talked a lot about agentic AI. The year before it was generative AI, and even before that we talked about machine learning. But what I feel has stayed the same in all of this is the balance, sometimes imbalance, between hype on one hand, we need to do AI, AI is awesome, we can do all these things, with all the doom and gloom and the big promises, and on the other hand, what is the actual impact? So I’m curious what you have seen there. Where are the challenges, and what works when we compare hype versus impact?

Ivo Strohhammer: I guess that’s similar for everyone when we talk to leaders about a new technology, and that has been the same in the last transformations as well. Everyone wants to be the first one. Everyone wants to have this magic box where you just push that red button and it solves all your problems. So there are big expectations in regard to impact, but most of the time we forget the people, right? Why should we do something? How should we do something? And the first thing, at least in our ecosystem, was that we were looking desperately for use cases. Where do you find these use cases? Just by asking someone, what’s your use case? What can you do, without knowing that they don’t really have a clue about AI either? Then it’s just a random list of ideas, right? So top down, we see that, and especially in the last year, it was growing: we need productivity, we need impact with AI. But on the general employee level we are still trying to find out ways. What can I do with it? How can it help me from an everyday perspective, into a process perspective, into new products? That’s a transformation where we really had to be careful with change management and the whole transformation support. So we started two and a half years ago with a program called AI by Design, which is basically nothing less than a huge community to foster enablement and create awareness, but also having all the experts, from a legal perspective, a compliance perspective, cybersecurity and sustainability, to answer certain questions together, to get a better understanding of what we can do with the new technology and where it really helps us. We tend to just sprinkle new technologies like AI over our process and hope that it gets us better results. But in certain areas we have to drastically rethink how the process should look, to really harvest these big gains.

Andreas Welsch: So many great parallels to what I’ve experienced in many projects, especially in larger organizations. We need to go on a fishing expedition. We need to collect use cases. One of my team members back then called it innovation theater. You say, give me your best use cases, and like you said, most of the employees don’t even know what to look for. What can the technology actually do? What are the opportunities, the challenges, the limitations? So they pour their heart out about the things they’ve always wanted to have fixed, but most of those things are not even solvable with AI. Another part isn’t solvable with the data you have. And then you end up with very few that actually make a meaningful impact. And to your point, on the other hand, the community is helpful to drive that awareness, so you can have these conversations about where it adds value and what we should actually do.

Ivo Strohhammer: Or even worse, right? On one side we have to change our mindset so that we are faster and experiment a lot, but meanwhile our experiments become the next feature of the next AI solution. Then the challenge comes: why do we spend money on experiments when that becomes a feature in half a year? So we want speed, but not too fast, because we don’t want to reinvent something which comes for free somewhere soon.

Andreas Welsch: That reminds me of the chat-with-your-PDF example, the early examples of startups and companies building something that everybody else then built into their product natively. So, you talked about this bottom-up approach. You talked about building a community through AI by Design. It sounds like it’s working really well, but knowing how large enterprises work, I could imagine there are also challenges in the approach. Looking back at two and a half years of your journey, what were some of the challenges that you didn’t expect going in, and how would you do that differently nowadays?

Ivo Strohhammer: The best approach was what we just did, right? It’s something new. No one really had the answers. There were tons of books out, but no real playbook for how we should approach this transformation. So we just started, and exactly one year later your book came around the corner and it really helped. What we did was start with a flexible, agile approach, gathering a community together, getting the champions network so that we know who is working on AI, and then really an empowered community, an open community where we ask for feedback, where we gave feedback and early access to tools so that we can learn. On a yearly basis we reinvented our transformation program. All the core team, with members from all different functions, came together for a day and said, okay, how should the next year look? Because in these fast-turning worlds it’s impossible to really drive or define something for the next five years. So, one-year cycles. At the beginning it was all fine and great, an awesome community, and after one and a half years the pressure for impact was bigger and bigger, and it was all about productivity. We know how hard it is to get from an idea to a productive use case. Experiments look fancy and good, but deployment is still harder than we think, and the pressure was then really high. There were multiple programs coming up, top-down programs, bottom-up programs, and different regions were running at different speeds. So it’s about keeping this whole herd together and fostering transparency about what we are doing, right?

Andreas Welsch: Right.

Andreas Welsch: And that mix, letting people experiment but also trying to standardize certain topics so as not to reinvent the wheel over and over again, and finding the balance. I think this is something that many of you in the audience have either experienced as well or are likely to experience soon, depending on your tenure in the role. I’ve seen this myself. You get about a year to build things up and to establish processes and governance and your community. But then, at the year or 18 months mark, the question comes: so what have you delivered?

Ivo Strohhammer: Yeah.

Andreas Welsch: And then it’s super important, if you’re early on in the journey, to focus from the very beginning on business impact and high-value scenarios. Yes, there have to be some quick wins so you can show some momentum. But I think you also need to be very clear about the resources you have, the resources you need, and the budgets, to be able to deliver something meaningful by the one or one and a half year mark, so you can continue with that momentum and stay on that trajectory. Otherwise, like you said, expectations are high, not only for the AI leader. Those who are asking you what you are delivering have even higher expectations raised towards them as well. So it’s not always easy to keep them in balance. You said it’s about building the community. It’s about sharing what’s possible. It’s also about hearing and listening: what are they doing with it, and how well is it working? What else was a good approach to get people engaged? There’s so much out there in terms of content and information, at least by now. How do you get people to actually use it and do something with it, and not just consume and absorb the information, say okay, fine, and go about their business the way they always have?

Ivo Strohhammer: There were two components that were really crucial in doing that. Within AI by Design we have a work stream called people enablement, and that has two directions. On one side, it’s telling the people why we need AI, why to use it, and giving them some guidance on how to use it in a safe way. The other is early access to playgrounds and secure tools to experience it. You can read all about it and watch videos about it, but you have to test it and experience it on your own. That sparks the next level of ideas. When I can write an email, how good is it with my documents? When I can write a document, how good is it when I combine this and this and this? That’s the only way we get to a certain quality level of use cases. We also have a learning platform, and if you type in AI at the moment, it gives you 6,000 results. That’s the obvious Netflix problem: you search for one hour for the right training and then you don’t have the time for the training. So we were curating lists of preferred trainings. Our cohort, our community, has what we call the role of AI champions. It comes with four levels. The first level is just being part of the community, getting access to the community material and the chance to communicate. The second level is theoretical knowledge, entry-level trainings to really get people onboarded. We said these are our proposals for trainings, but if you have done something else, come to us. We don’t want to spend ridiculous time doing it over and over again, but there is a certain baseline of knowledge, theoretically about AI but also about fostering AI use cases, the ideation. The third level of the AI champion is practical application. I either worked on use cases, or I worked on ideation workshops, or I facilitated trainings, to grow that AI champion score. And the fourth level, basically in connection with a talent community, is called the AI captain. That’s a single point of contact for a division, business unit or region. So we have a kind of pyramid, with 10 sub-communities in the US, sub-communities in Germany, a sub-community for a particular business unit, and all of it is really hands-on experience as fast as possible, so that they can get the word out, we can share the experiences and we can build on each other.

Andreas Welsch: I really like the idea of the tiered or pyramid system. Also what you shared, basic access for anybody who can read and consume the information, but then, based on the additional qualifications and skills you gain, you can advance and also become a trainer for others. I think that’s great, especially the part about getting hands-on. Typically when I give trainings, I’m always reminded of the bell curve. You have some who are super advanced. You have the vast majority somewhere in the middle saying, “Yeah, I’ve done a few things. I know how to write an email with AI. You don’t need to tell me about that.” But there are other things, like prompting and advanced techniques. I didn’t know it could do that. I didn’t know it could compare documents, or something like that, right? The vast majority. And then toward the end of the bell curve it’s, “Oh, I’m not really sure how to use this. What happens with my data? What kind of data can I put in there?” and so on. So I think it’s really important to meet people where they are, but also give them an opportunity to grow in their learning journey.

Ivo Strohhammer: And to give them a secure opportunity to grow, because people are using it. It doesn’t matter if you offer a secure copilot or not. People are using tools, and as soon as we don’t offer a secure way to the technology, they use what they find on the internet, and we all know how secure that can be, right? That’s the governance nightmare of the enthusiasts, the early adopters. They’re running with the technology, and the technology is so approachable for everyone that we have to be fast, as a company or as a larger network, to offer a secure and safe version of it.

Andreas Welsch: I think it’s super important that you need to do something whether you want to or not, because your team members, your employees, are going to do it. They’re going to find ways to use it. They’re not being bogged down by rules or by clamping down access to most of the tools, or by a wait-and-see approach. They’re running with it, like you said. So it’s inevitable in that sense. As much as I don’t like the narrative of it’s here and it’s inevitable, you see this in the usage statistics and in many reports as well. Now there’s one point I wanted to come back to. You said that right now a lot of the focus is on productivity, and I see that too. It’s about how we get people off the public version of ChatGPT onto something that is governed, something that we either have licensed or have more control over. Lots of large enterprises host their own LLM playground or model garden, or some way to experiment in a sandbox. But a lot of the focus I see right now is on personal productivity. When I wrote the book, there was a lot more focus on operational efficiency and strategic differentiation. How do we use AI in our business function to do more with less, to get better insights, to engage differently with our customers? But nowadays so much of the conversation is about how I write emails and reports, how I summarize things, how I transcribe meeting minutes, and what have you. Where are you seeing this right now, and where are you seeing this going in 2026, for example?

Ivo Strohhammer: We segregated AI into three buckets. One is everyday AI. That’s the personal use of everyone. The second one is process AI.

Andreas Welsch: Mhm.

Ivo Strohhammer: So the big processes, end to end, standardized. And the third one is new AI products and new AI services. In 2025, the focus was on that whole process AI part. How do we really make it better and faster, with productivity on that level?

Andreas Welsch: Mhm.

Ivo Strohhammer: Because let’s face it, personal productivity is a really difficult thing to measure. How are you as an employee rewarded today when you’re faster with your work? Normally with, okay, let’s do this as well, right?

Andreas Welsch: Or more work, right? When I found out as a manager that you were more efficient, I gave you more work because you had more time.

Ivo Strohhammer: Exactly. Exactly. Right. So people are using it but don’t really tell.

Andreas Welsch: Yes.

Ivo Strohhammer: How do you measure that adoption? But for me personally, I think that is the key driver for the whole success. Because on the process AI side, as I mentioned before, does sprinkling AI over the customer service process make that big leap? No. It gives us a couple of percentage points of productivity. We have to drastically rethink how customer service could look with a new technology on board, to then completely reshape these processes. Is a big international company ready for that? From a risk awareness perspective, probably not, right? But personal productivity is the key for everything. If people are motivated, if people are driven by that intrinsic drive to become something, and if we are then able to reward these people, well done, spend your day at the beach if you’ve finished your work on Friday morning, for example, then that fosters everything else. That fosters better ideas for the processes, that keeps us open for creativity, for new products, for new services. So I hope that shift goes a little bit more to the individual, to that personal usage, to that personal productivity.

Andreas Welsch: So now you’ve obviously seen this in and out in large enterprises. I know you said you’ve been a consultant prior to that, so you’ve seen a lot. How are you now bringing this to small and medium-sized businesses, and what do they need to do? How do they need to think about this so as not to repeat the same mistakes that we’ve all made in larger enterprises, but benefit from those learnings and invest the limited capital and resources that they have effectively to get some good output? What’s the net for small and medium-sized businesses?

Ivo Strohhammer: It’s a great question, because they are lagging behind time-wise, right? That’s the big advantage I have: when I send an email to one of the tech companies with a Siemens email address, I get a meeting the next day. If you’re coming with your email address from your small or medium enterprise, best case you get a sales rep on a call in a couple of weeks. So what we said, especially now in our area, is: let’s create a physical lab with open doors for everyone, and we work on a triple A framework. We want to create awareness, we want to foster ability, and we want to really drive adoption of use cases. First, leaders from smaller companies should know what they can do with the technology and, more importantly, what they can’t do, because it’s not that magic box so far, right? The second thing is empowering the workforce to use it, over and over again. And the third one is really harvesting the productivity gains. Smaller companies don’t have the luxury of having their own IT department, AI specialists and so on. So as a community-driven approach, we want to help each other, supporting with standards, supporting with industry-wide solutions, supporting with out-of-the-box trainings, so that we can make the best out of that limited time, limited capacity and limited functions, not reinventing approaches over and over again, but learning from each other.

Andreas Welsch: So awareness: what’s AI, how does it work, what do you need to know? Ability: being able to build it, and then actually build it, right? Apply it.

Ivo Strohhammer: Yes.

Andreas Welsch: I think that’s awesome. I love things in threes, so awareness, ability, application makes perfect sense to me. Now Ivo, we’re coming close to the end of the show, and I was wondering if you can summarize the key three takeaways for our audience today when it comes to moving from pilots to programs and how to make AI stick.

Ivo Strohhammer: First of all, don’t forget your employees. That’s the most powerful resource we have, and it’s the biggest democratization of technology and capabilities we have ever seen. So use your employees, train them, teach them, give them the possibilities. Second, find a way between creating productivity but also fostering experimentation. Let’s face it, experimentation will create a lot of similar use cases, but that’s needed to learn, right? So don’t just focus on the high-productivity use cases. Find a balance: let them do something, probably in parallel, and then find the right way to standardize. And the third one, keep an open eye, right? It’s so fast-turning. We don’t know what’s tomorrow, what’s in a month, what’s in a year. I’m far away from defining strategies right now. It’s really about having that agile mindset of being open to whatever comes. Test everything. If it sticks, fine, roll it out. If not, fail fast and move on.

Andreas Welsch: I love that. I think that’s so important when things are moving so quickly, but there’s so much already there to learn about, to experiment with and to apply. So Ivo, thank you so much for sharing your experience and your expertise with us. It was great having you on the show finally. Thanks for all your support.

Ivo Strohhammer: You’re welcome.

Andreas Welsch: And best of luck with the new endeavor and the new venture.

Ivo Strohhammer: Thank you. Thank you. It was a pleasure. And let’s stay tuned and continue that great exchange.

Andreas Welsch: Fantastic. All right, thank you, folks in the audience. See you next time for another episode of What’s the Buzz. Bye-bye.

Strohhammer said his team started with a flexible, agile approach and a community called AI by Design. It combines enablement and awareness with experts from legal, compliance, cybersecurity and sustainability. The core team reinvents the program every year, because defining it for five years is impossible in a fast-moving field.

Strohhammer described four levels. The first is joining the community and its material. The second is entry-level theoretical training. The third is practical application, such as working on use cases, running ideation workshops or facilitating trainings. The fourth is the AI captain, a single point of contact for a division, business unit or region.

Strohhammer said people use AI tools whether or not a secure copilot is offered. Without a secure option, they use what they find on the internet, which he called a governance nightmare. Companies have to be fast in offering a safe version.

Strohhammer described a triple A framework. Awareness means leaders know what AI can and cannot do. Ability means empowering the workforce to use it. Adoption means harvesting the productivity gains. He also recommended a community-driven approach with shared standards and trainings, so smaller companies do not reinvent approaches.

Strohhammer said personal productivity is hard to measure but is the key driver of success. When people are motivated and rewarded, it leads to better ideas for processes and keeps teams open to creativity, new products and new services.