Episode 78

Olaf Maecker on Building and Leading Agile Data & AI Teams for 2025 and Beyond

TIM: We're live on the Alupa Objective Hiring Show, the first one for 2025. I'm pumped to speak today to Olaf. Welcome Olaf.

OLAF: Yeah, thank you for having me.

TIM: it's absolutely our pleasure. And yeah, welcome to 2025. It's exciting. And I would love to start with getting your thoughts on team building in particular in data and AI and, you know, from your experience in building teams. How do you do it successfully? What are some pitfalls to avoid? How do you build a really strong data and AI team?

OLAF: Thanks, Tim. I think this is also the most important questions question when it comes to, to building a data and AI team, what is successful to contribute to the company, this on the one hand side, of course since years. a data topic around and the company company is realizing how we need people who can do something with our data. But the challenge is always how to implement it on the one hand side on an organization. But the more important and most important topic is to find the right, the right team set up. It is I compare this always a little bit to, a soccer team. So it doesn't, you can't win the game with only, Goalkeepers or frontrunners, you need in the team who cover the different stages, what you have in the data team, sub elements, like more technical driven, more usually the data engineering part. So building the backbone. building the data pipeline, make it possible that the scalable connect the relevant data sources. On the other side, you need to send the people who do more with the data. So building data products, AI products, last but not least, also not to forget unfortunately, I would say a little bit the most important why it's unfortunately, because from a complexity point of view is building this whole data pipelines and data products is very difficult, complex. bringing this then to the business, this is a key success factor. You have, can have the best people in your team, the best data scientists, the best data engineers. But when you're not able to bring this to the business, you are not able to unchain this power of the team. So what is important there? It is about to position your team, the different disciplines. So people who knows. How to build a data warehouse, how to leverage AI capabilities, large language models, when you work with a hyperscaler like we do, you use the service in a way end to end to contribute. And of course you need people who Navigate the organization communicators. I call it data navigators. Other call them data stewards, people who are able to talk to the business in the language of the business. And when you are able to build this data team in this way, you will be also very successful if you're not able to the glue together between this rules and also the glue in the direction of The business, it will be very, very difficult. So it's in the end, again, like a soccer team, you have to position your people in the right way to be successful to win the championship.

TIM: Well you're talking to someone who loves football and loves a football based analogy. So I think we can continue this train of thought a little bit more. So, it sounds like then, you're basically saying you need to have this portfolio of skills. As you say, you don't just want 11 people great at keeping the ball out of the goals or 11 amazing goal scorers with no one able to defend. Is it then a case of just filling in those positions? So I need a left wing. I need a right wing. I need a data engineer, the data scientist, or is it more at a skill level that you would think about it. I need communication skills. I need visualization skills. I need data pipeline skills. Like, do you think of it at a role level or skill level? Or is it a bit of both?

OLAF: in the end it is bo both. So I implemented always skill metrics in combination. So when I hire people, it is important that the foundation is there. And for me, it's not in detail about, are you able to use the services from Google or AWS or something like this? It is about, are you have experience with this? Are you able to learn new things? And you can learn this kind of skill, but it's always about a more soft skill topic. So I have here five, five important topics. What brings the relevant topics together? For me, it's like a hygiene factor that you're able to code, that you are able to know how to use simple statistics, how to use AI, how to use large language models, but you need this kind of sense. with the people to, to bring this to success. And this as I said, a little bit, most of the soft goods skills that is about ownership. So when in the end, not a data AI relevant topics are always important, but we need people who have ownership behind this topics and it's about can do culture. So also think about. How you can make it possible with this, what you have, because in the, in the we are living, the data is never perfect. unfortunately the case that you have to build something around. You need proactiveness, you need business understanding, you need a navigator sense, I would say. So to help them to recommend, but last but not least, also the, the topic to burn for data. So it is a combination of both. Yes, you need the skills. You need the right techniques. You have to be always up to date when it comes to data topics. The AI story of the last 18, 24 months showed this. That was so fast developing. was then always not about that. I know exactly what this company or that company can do. It is important that I know that it's possible and then find the right combination. So it is also about how to combine it the right way. So yes, you need people a, from a character skill basis that can do stuff a good way. On the other side, of course, they need also this kind of curious, have hunger to do something with the data. I call this always, I have to burn for data. I have to see the fire in the eyes of the people that they say, okay, data is a topic, AI is a topic I want to do. And when this is there and they have this foundation of skills, is possible.

TIM: And so basically we're saying we need this portfolio of skills, the soft skills, the technical skills. And so that's all about you're finding the right people. Are there other elements though in making that team successful? So that's like the recruitment side, like assembling the all star team. But what's the difference between just having a whole bunch of amazing players and a coherent team, I guess is one way of putting it.

OLAF: Yeah. Hiring is, is the crucial part. That is when you do there a failure, hiring, hire the wrong person. is very difficult to fix this, you know so it is important to find the right talents and there I look more for attitudes, less for skills. Again, there is a threshold that has to be there as I have to know how data works, how statistic works, how econometrics works, but It is not so much important if they have so much knowledge in Google services, AWS, ATC, it is about that. They have the right attitude and to find those people. This is very, very difficult. because. When you look, see only the application, you look on the CV, usually it sounds good. and you see also certificates, what looks good. But if the person is able to use it really in a sense of business and contribute to the business, or is it only like, yeah, I learned this and I know the terms. But how to contribute? I'm not able to do this. You have to find out in the hiring process. If you fail there, you will invest a lot in the onboarding. and after six months, eight months, 12 months, we realized, Oh, this candidate is not fitting. It is not the scorer, the gold scorer. It's not able to have. Also when it looks on paper, great, but you need a person who are able to use the skills in a, in a way. What is helping the business?

TIM: I can think of quite a few football managers over time who have basically said, look, end of the day, 90 percent of its recruitment. You just got to focus on that. Getting the right people into the team. That's really where you're going to win and lose. And that sounds like you're saying the same thing, but for for data data teams. What about incorporating a new player, a new person into your team? Like, how do you help them be as successful as possible to really contribute as much as they can to the business?

OLAF: Yeah. There are two, two things. One thing is of course you need the hiring process. A very good onboarding that you're not say to a person, Hey, Olaf, you are now new to the team. Good luck. You need a good onboarding. That's a person has the chance to learn about the organization very quick that he or she can contribute. The second topic is. The existing team. Of course, everybody has to start and say, okay, I'm starting in an organization. I have to build a data team. You have to build this and bring step by step this kind of team spirit in. But when you are successful, you have a good running team you created a spirit of data burning for data and a spirit of a sense of team that they are not Thanks, Carl. only there for the money. They are there to do cool stuff with AI and data. Then you have already solved a lot because when a new person can now in the person will be in a positive way infected the spirit. And then this is a safe dynamic, a safe dynamic loop what will come. So most important is build a data team and AI team. What is creating a very positive spirit. They're very motivated and doing cool stuff with data and have this. spirit of, Hey, we are the coolest team in the world on the one hand side, but we have a lot of fun in working with us. And when you throw a new player in to the game, he will, she will be infected in a positive way. So, okay. I'm part of the school group. I want to deliver more. And I am very interested to, to deliver cutting edge data and AI technologies. And I think this is a little bit the trick behind. So good onboarding that's clear, but when you have a good data team already around with a good spirit, then the rest is automatically done, I would say. And it is consumed by a team, by the team and the person can very quick contribute to the business.

TIM: The natural next question, I think that anyone would be asking is how then do you create that amazing team spirit in the first place?

OLAF: Yeah, it is about creating. An environment on the one hand side that everybody can bring himself or herself in so feeling that everybody has a feeling and also in reality that they can see I can contribute and of course we have also. in the team. We have juniors with middle level seniors with working students, I always say to the people, Hey, we sit together on a round table. And if you have a good argument, a good topic, it is not important. If you know, the senior data scientists. working for 10 or 15 years, or you're maybe the working student and bring something fresh in. So also to reduce the topic, like, okay, I am only allowed to speak if my mentor or senior person invited me, it is about, we are one team. And when I can contribute, I can contribute. And every topic, what is said, It is not see as a, as a criticism to them to my personal work. It's about, okay, can we improve together and by this content topic? What's the person just mentioned? Can we improve as a team and be better and deliver best? I think this is a of, of the topic to invite everybody to have a safe space where you can say everything. course, you should think about this, bring this argument on, but in the end, it's count this, what you're saying, and the idea you have, and it's not counting which hierarchy level you have behind here. this is, this is, I think, an important topic to invite then everybody to contribute. On the other side it is. It is important have projects where the people can say, I can invent something. I get the ownership of this. I have something, I develop a new data product or the sub element of an AI functionality, what really brings, this ahead. And I'm responsible for this and not like I built here something and then I go to my manager. He gets there's a positive bonus or it is about them to say, okay, this is what I can deliver to this. And when you combine this, have this feeling that coming together as a team to contribute to something, what is amazing and put this. important projects in place that everybody can contribute. I think then you create automatically this spirit.

TIM: Yeah, it's such a great summary. And so it's almost the way I'm thinking about it now is give a person a chance to create their own little thing. It's almost like doing a startup in a way, but just on a micro scale. And whenever you. Do a startup, you have this, it's like, it's my baby, you know, it really matters to me. This is what kind of my creation, I've managed to create this from nothing. And so I guess it's almost a similar kind of thing again, maybe on a smaller scale. But people would be naturally more bought into things that they've created themselves. Does that make sense?

OLAF: sense. So to compare this a startup, it is, it is a perfect comparison. It is a little bit like a company in a company, of course. And, but it is like to have the startup feeling that is a very good summary. That's a people say, I'm part of something what is important and amazing. Also, maybe you work for a big corporate organization, but inside this corporate organization, you built a data and AI hub and I'm part of that's my baby. I can contribute. This is my footprint. I can deliver. And when you create this feeling in the team, the people not asking about, Oh, is it already 5 p. m. I have to go home. It is about, Hey, I built a cool project, a cool product. I want to do my best and deliver this in the best quality. It's like a startup. It's good comparison. TIM: what about in terms of like a shared vision for the team or having clear, a clear team goal or metrics the team are responsible for? Or is that also part of it, do you think?

OLAF: so we always, when I build a data team, we create a team vision. Of course the vision has to fit hundred percent in the business vision, in the company vision what is behind. But we say we are navigator for data driven business success. So that's the claim I have currently in. And Come again, again, also know workshops as a first slide. Say, okay, this is what we stand for. We are not waiting for requirements. We are not a traditional IT organization. We are more business orientated and say, okay, we have the responsibility and the organization to help the other parts of the organization to unchain the power of data analytics. And now what is difficult? The other parts, you usually don't know what is possible with data. Okay. then you need this people, the persons in the organization who say, Hey, there's one more thing you can do. Maybe say it's us asking, Hey, Olaf, is it possible to, to increase our conversion rate for up and cross selling and say, yes, of course I can build you for you a next best offer system, but I can also help you. That is not only about deliver one time to use a relevant customer. You should do up and cross selling. It is also about deliver a system you can do it automatically over multiple channels to use the data you have to identify what is relevant to bring this into the customer. And maybe I can also deliver you some insights that you understand your customer segments better. And this is examples where I would say it is not only about to deliver the exact requests. That is important to do so, but it's also think about from a data perspective and AI perspective, what is possible, understand the question, what the colleague want to, to solve here. And when you are able to adapt this and say, okay, I deliver this to you, but by the way, you can also do this. And this was AI. Maybe it's interesting for you. Let's do a pilot together. This is something what bring, bring you ahead. And this is a principle of being a data navigator to help the captain on the ship to come to the port with, with, with the cargo ship and not say, okay, you are alone and I take no over or something like this is about teamwork in the organization itself. But for this, you need this data, data navigator principle.

TIM: And coming back to some of those kind of soft skills that you mentioned, that's why the proactiveness, then it's so essential because I could imagine a lot of analysts and some teams might get a request. That's a juror ticket. They do the juror ticket. They build the dashboard job done kind of thing. But that's obviously not what you're after because you want something a bit more value, add a bit more, as you say, business focused. And then I guess that's also the key. in terms of understanding your data, understanding what's possible with data and understanding the business. If you can somehow connect all those dots, then I think you can become very valuable or your team can be very valuable in a way that you can't, if you like just know the business, you don't know the data or you know the data, but you don't understand what can be done with it. Or you know the data, but you don't know the business. Like it's, it's like having the combination of those things, I think suddenly is gives you massive leverage and. makes you much more powerful or valuable than just knowing one of those areas, I feel anyway.

OLAF: Yes. So it is. Of course, important to organize your work. And yes, we're using Jarrah, of course, but it's exactly as you described. It's not like I sit in my chair and wait that somebody asked me exact the question. It is about exchange with the business. And this is always as good as comes to data to understand what is the real business question. What is, what is the topic the colleague wants to solve? And again, the colleague sometimes not not really know what is possible with AI, what's with the data we have in the organization and the different tool sets. I have to help the person to buy by being also proactive to understand the question and try to find out what topic you want to solve. And it's not about like, Hey, I need a list of 10, 000 customers. I can do an upset. So, okay. What do you want to solve? Yeah, I want to maximize the customer value on. So is it about a one time sale or is it about to over time do again up and cross selling to maximize the customer value? And to implement here something and to understand this, you need this, this curiosity of the of the, of the data person itself to ask exactly these questions and not say, okay, this is a question I answered. It's about, okay, what is really behind and you have to be a real data scientist, a person, explorer in the end to, to step into, into the new country and say, okay, it's not only to set my foot into the country. It is also about. to find out what is here, what's in the jungle find out how can I help better? And this is only possible when you have a proactiveness. And yeah, save curiosity here behind, say, okay, I want to understand once a, what's a person want to solve.

TIM: Yeah, so it's a curiosity, a proactiveness, an empathy as well probably in just wanting to help someone solve their actual problem, which may not necessarily be exactly what they've asked for if they've Kind of simplified it. They've gone for a solution, but you're always better than this. And the problem, I guess it's about also asking questions, like just asking good questions to kind of dig a little bit deeper. Anyone in sales would do this. I think any good analysts should also be able to ask those questions. May is one for you. Have you noticed, does that ability to ask good questions? Is that something you generally get with experience? So would it be you sort of graduate analyst who would just. quick deliver the first thing that matches what the person's asked for. Whereas maybe someone is a lead analyst with five or 10 years experience might be thinking, hang on, like, I know what they're really asking here. And it's something slightly different to what they've said. I'll know off, I'll ask them these questions just to clarify, like, is that just something you kind of get with experience?

OLAF: Yeah. Yeah. This is the important thing when we new project starts to ask the right questions. And usually you need a more senior person because the junior person knows the skills, maybe how to code, how to implement a machine learning model etc. but this things between the code and the result. It's about to understand what is the real question. Otherwise you do the topic what is asked for, and then you realize in the end of project. Oh, I feel result. But my expectation was different. I saw this a lot of times in data teams in my consulting time. They did a very good job from a technical perspective and solved and answered a question, but they understood the question the wrong way because they said, I thought the colleague want to answer this and that. you need a person who minimum do Recheck off the question. Also, when you hear the question to say, okay, I understand correctly. You want to solve this and that. And you do this because you want to maximize the customer value when you come back to the example. And when then he agrees that yes, that is a topic. Then it's good because then I did the recheck and can, can go to back to my desk, talk with the team, what to do, translate this from a business language to a data language, if it not correct, then you will get an answer. What helps you to understand this better. That is in the end also the rule to ask the right questions that the senior person who is in contact for, for project start, understand it. Exactly what is to do, but this person is also trying to translate this to a to a language what the development team can solve himself when it's a smaller project or when we have the bigger project where maybe two, three, four people working parallel on this, that everybody has the same understanding what we want to solve. So data analytics, AI is, has a lot to do with creativity. You have to, and have the, yeah, the possibility that is maybe a positive thing, take so much decisions during the analytics process. But for this, you have to understand the question right away and yeah, it starts with the right questions in the, in the kickoff phase.

TIM: I'd be interested to hear your thoughts on where these roles and capabilities are going to go. Given how quickly large language models are advancing. So you kind of laid out a few different types of challenges that analysts might be exposed to. One is a slightly more transactional. I need X, I need Y. I need this report, I need this dashboard. I'm trying to understand this specific thing. Other things become more complex and require more creativity, require a bit more digging, require a bit more thinking. Do you think we'll get to a point where those business people For their kind of meat and bread, meat and bread, bread and butter kind of request, they are going through some kind of LLN layer that sits on top of the warehouse or on top of the dashboards to help them convert their English language query to a SQL query or whatever. And that they would use some kind of tool there to get those basic answers. And then the analysts or data scientists might end up being leveraged for more complex, deep dives or models or something beyond just simple BI requests. What, where do you think things are going to go?

OLAF: Yeah. So I feel a clear prediction. You look on the data analytics process and you have the descriptive analytics, predictive analytics, and the other stages. I see already the trend, but my prediction is in one to two years latest, that the basic questions when it comes to descriptive analytics will take over by an AI agent. Of course, you have to build the system in your organization, but If the set up you not longer need this basic functions in the end, it's about why I call it basic functions. When I look on a data team, our goal is always to build multivariable models. What helps us to deliver more? success for the company, but we are blocked usually by simple requests. Like, yeah, what is the sales performance of the last 12 months in the country? X, Y, or Z. And then the person go to the data, find this out. But this is an easy request. It is about a secret query. maybe to do a cross table and something would say, yeah, that's important. That is very important for this. area of the organization, this is blocking us to build data products and AI products. So I have a clear prediction that more and more we will see software, AI agents, where they can directly talk to. That a salesperson can ask a ai agent, say, Hey, can you show me the sales in country X of the last 12 months and correlates this with a weather forecast, something like this, and then you will get a result. The foundation, and this is since 60 years, since 100 years. You need the right data source, bad data creates also bad results. Also when you put an LLM on top. So it's an LLM can also only give the answer you see in the data. But when this is there, we see more and more that this rule is less Relevant to put a human on top the machine will take over you.

TIM: Yeah, I assume that's where it's going to go as well. And you kind of imagine there's going to be some teething issues. Maybe the companies that adopt these technologies a bit sooner are going to have to deal with maybe a little bit less. Accuracy, but it's also clearly a trade off because I mean, the number of questions that, as in your example, a salesperson might want answered any day is almost unlimited. I'm sure at the moment, they're kind of just focusing on the most important things they need to know. And they're the request they're firing off to the BI team because they know it's an effort to do and what have you. So if we had some automated way to do that. Surely we're better off being able to answer 100 questions with 90 percent accuracy than I don't know, three questions with 95 percent accuracy. That's that's my view anyway. What do you reckon?

OLAF: Yeah, it is You need always of course a person who can dig a little bit deeper when something new comes up that is That is 100 percent sure. I would say it is more about to hire the right people in a way that is not only just the person who do in report, it is about a person who are able to develop AI agents who can ask answer the standard questions, but who is also there to answer new questions. And help them to build a project. So I want to understand the new customer segment or what are the drivers for our new product. There you need a person, a human who are able to analyze, look on the data, derive implications and then coming back and say, okay, I can help you. With this in this insight, I'm here, your data navigator to help you to reach, reach the goal and bring this in. And of course, this question has sent to be implemented in the knowledge base at the agent in the future can answer you. So it is about, yes, new topics, new questions. The human has to cover this and with expertise to bring this in, but it is also about to train the agents for the more easy descriptive analytics things. When the standard questions came up, to cover this,

TIM: Yeah, it might end up being a kind of triaging based approach, like you might see customer support bots at the moment where you initially get a bot that's using, as you said, the support documentation and maybe the website to answer questions, but then if the question is too complicated and it kind of gets pushed to the human, maybe it'll be a similar situation with like an analytics bot sitting on top of your warehouse, interact with it. And if it's just too complicated or it doesn't have access to the data set you're asking it to access, then it deviates to an analyst. To dig into it in more detail. That would be really fascinating. Actually, if you had some way to capture every person in a business is analytics question all in one place and that's then being improved on the knowledge base all the time. So you kind of got this demand for insights being captured as opposed to in emails and slack threads and what have you. It was all kind of sitting through some integrated BI AI tool that would be kind of cool, I reckon.

OLAF: yeah, this example with customer services is very good because I think this is. the hot topic in a lot of organizations. Also, we are working on this. Because this is close to a problem, what an AI agent can handle, especially when you come back to standard questions. When you say 80 percent of the topics. ask again and again, again, then the agent can take over, but you need always the flexibility of the systems that a human can step in. There are then more important topics or more complex topics for the agent is not able. And the And the only thing is that the AI agent has to. Understand. Hey, there's a question. I don't have the knowledge of my knowledge base. I have to hand this over to a human colleague. And then, then, then you have this solution. And of course, in the future, the AI agent will learn, say, okay, there was a question. I handed it over to my human colleague, but I saw the answer. And now I learned from this. Next time I can answer myself.

TIM: We're basically thinking that the prescriptive analytics might be one of the first bits that gets automated or enhanced by large language models. Are there other aspects of data roles or specific data roles that you think are going to be quite drastically changed in the next, let's say, one or two years? And, and yeah, what are your thoughts there?

OLAF: Yeah. Yeah. So I think you have also a gradual steps in the different rules. So one is when you come to the descriptive analytics, we say, okay, that we will have see also more replacements. the other side, you will see in the other rules. More a combination, adding AI as a superpower to your own own topic. And yeah, the quote is out in the world since months and everybody knows this, but I can only repeat this here. AI will not replace the data in the AI team itself, but you will be replaced by people using AI. In the daily business, and I can see it on my on my own daily work. Of course, I'm as a chief data officer, a data leader, more organizing the teams and motivating the teams, less code and do my stuff on my own. But since I use AI, I boost my efficiency so much. It is about building the right concepts. Checking specific things when it comes in, having summaries, having meeting minutes. It's helped me to organize and be faster in my work of the day and more efficient. And of course, also have the right quality check because everybody knows when you write a concept or a long email to do the formatting in the end that it sounds very good. It takes a lot of time. Nowadays I use my, my, my AI agent, my chatbot to say, okay, I have this email draft. Please. Do it in a nice way. I want to send it to the CEO and then I get another draft and I have only to do five minutes rest work and then I can send it out. So I'm so much more quicker. And when you look now to the data people who do the real development, that is amazing what you can see there because it is not only about using AI in the daily business, like writing emails, it is about the coding. the interesting stuff we did it in 24 as a pilot with a coding assistant was, it is not only to create new, good code, it was very important for the people to understand the code. And it's not about that. They're not able to understand it. It's about, okay, I put this in and I want to know exactly what the code is doing and maybe how I can do it better and do in parallel quality checks that not I miss comma or important statement somewhere because I have a long list of code to see what I can do. That is what's interesting to see. It is that the people by the daily business use the AI coding system differently that I expect they do. And it was very surprised, but they add this to the daily routine this basis, they are so much faster and deliver in higher quality.

TIM: What are you thinking now in terms of? The skills that you're going to require your people to have just because large language models are improving so quickly. And as you say, they're kind of changing the workflow quite quickly. It would not be surprising to me if in a year, the idea of writing code from scratch as a human would be ridiculous because. It's going to be this interface where the LLM is doing on your behalf. Maybe you're still reviewing and maybe you still have visibility. Maybe there's still existing bases you have to dip into. But like, surely we're getting to the point where humans writing code from scratch is less and less common. I don't think that's a controversial statement. So that may be like hiring for SQL skills or Python skills for then. How, what it is to be a good engineer and anything that revolve in coding might become redundant, basically is what I'm trying to get at. How do you think about that changing blend of skills and other certain things that now might be more important, certain things that might be less important. Some people I've spoken to have mentioned like adaptability is, is one thing they're thinking about is because well, things are changing so quickly. I'm not even going to focus on. Just specific technical skills. I just want to make sure someone can really learn quickly. You can adapt, has that mindset. So some people are going down that route. What are your thoughts? Like, what are you, what are you hiring for over there?

OLAF: Yeah this is, this is very important because this is changing in, in the past, I look hire people who were very good in the technical skill itself to able to create code from scratch to know a lot of SQL statements and have smart things to, to implement machine learning models in Python ETC. This is of course still important. It is not about that is off the table. But it's more and more important to have the things what brings the different components together. the glue behind here, it is about the ability to learn new things, to adapt to a new environment. So the, our area is developing so fast over the last 24 months. And the speed will still be there. is not about that. I know this LLM exactly a hundred percent because the next LLM will be out in six weeks. It is about the ability to adapt new topics, but principles. I try, especially for the, my senior people and my direct reports ahead of to establish a principle, following principles. In doing the work and not say you need exactly this tool. Of course, we have a tool set, we have a tech stack, what we're following, but due to the quick what we have, it is important that people are flexible to adapt, open to look to, to, to go to maybe another tool, not say I stick here to my Coding language what I learned in university or something. It's about okay. What can I do better because the large language models which are out and the tools which I will help you to do it anyway. It is about the principle how to do it. And if the language, the coding language is indifferent, it is not so important because the tool itself will deliver the right code for you when you can explain you what you want to do. And it's about more holistic thinking about the data and AI process, what you want to do less about the skill. I should of course, quality checks. I should be able to read the code itself, but it's more about what is a bit between this, what I want to deliver and do the coding and write, ask the right question to the to the tool, instead of being the best person in Fortran or something like this.

TIM: Right. And so those principles then are things you can kind of rely on because they're not changing every day. They're things that you can always look back on. It's almost like having a framework, I guess, that you can sort of orient yourself within. And getting back to the football analogy, I know that's one big thing. The Tottenham Hotspur manager goes on and on and on about. The Australian, Posta Koglu, he's always about, he has his philosophy of how he plays. That is never changing. He would rather die than change his philosophy. But, that doesn't mean he doesn't change the players, or tweak the formation, or change the style a little bit, or particular tactics within a game. But he's got this overall framework that he always uses. I feel like it's kind of similar to that then and maybe that's something people can that would help people, I think, in an environment that's changing so quickly, they can become quite stressful. You think like, what is even going on now? These tools can do anything. If you have that, the set of principles to fall back on that is almost de stressing in a way as well, I feel.

OLAF: Exactly. It is it is a good compression with Tottenham. So it is, it is about to have the philosophy behind what is more a little bit on a higher level important. That's okay. That are the boundaries, what I have, my guiding principles. That means not that I not change the way how I, how I play. I will do always the same, but it is not about, okay, do we have more the focus on, on, on the one player or the other is about the right combination doing this because when you follow the principles itself, you will be always successful. you have, of course, the right principles in place, you have to develop in a way that has a holistic view on this, what you want to achieve, but you need in this kind of environment, only the people who are flexible enough to say, I can adjust and adapt to the current situation, what I have. I've always in mind, I want to follow this three or four principles, what I established and importance, of course, also when you have this kind of principle point of view, that's a communicating way that everybody's understands this and can understand this, how this works and follow the principles, but the, the, the leaders for two on the one hand side, implement this principles on the other side to translate the principles that everybody know what does it mean for the daily business? How can I use the daily business in my daily work? And not like, okay, now I have this cool text philosopher who say I have to do this, this and this and have to follow. But my, my problem I have today is looks like this. How can I do this? So you need also a little bit of mentoring that people can adapt. But in the end, it's about the following the principles. And when you do this, you are flexible enough to adjust to new situations and then you will be successful.

TIM: One thing that struck me. When you're just talking about that then was, I wonder if because this technology shift is so profound and so big, which on one hand makes it hard to adapt just because things are changing so quickly and such, such a great magnitude. But I wonder because of that, then you're less likely to have those kind of people who are stuck in their ways and insist on doing things in the old fashioned way. Because how, how could they in the face of the. biggest hype ever. It would be like refusing to use a car in 1940 and just saying you want to still use a horse. Like, why, why would you do that? If it's so obviously better to use a car. So have you, like, have you, over the past couple of years, have you noticed anyone who was completely reluctant to engage with AI at all? Like even now?

OLAF: Yes. It is of course, not a bigger part of the people, but you see especially. who sit in a comfort zone that when now something new came in, they say first, yeah, but what I'm doing, it's working. And this is an independent from a data organization. You see this also always in, in other parts of the organization, when you have people who work a long time on one level in which one role, they pushing back changes. And this is something where you have to do two things. One thing is of course to give a helping hand and everybody has to understand we live in a world where long life learning is important. Like the university is always saying, but the same is important for a company. You have always to, to go one step ahead. Otherwise you will be the wrong train and yeah. In the dead end, I would say on the other side, has always to do with the person itself. They need this motivation, this curiosity to be an explorer, as I said in the beginning, say, okay, I have to explore something new. That means also I have to adapt new tools. And this kind of change is is new stage of a transformation of an organization. We have the digital transformation, but still ongoing. But I think the majority of, employees. Digital, not everybody, but you would say, Hey, in general, a company's digital. I see a lot of this when we came out to AI, we have the next cycle of transformation. And it's in the end, also my job and the job of my, my, my direct reports to transform this giving a helping hand. But yes, there are still people who say, no, I tried to do it like in the nineties. But they will be not successful because they will be slower. They will be not good enough to deliver a good quality project because it takes so much longer. It's not about the expertise. They have the expertise, it's about to be open for new things. And this, and then coming back to the hiring, is something we as leaders, and when we talk also with the HR department, it's important to identify this kind of skill, this kind of attitude, I would say it more in this way, that a person is always there and open minded and open to learn new things and not say, Hey, I have my PhD, I'm now final in my, my education. And for the next 30 years. I will only do this. And that, that is not the case. We have to identify them in the hierarchies. And this is difficult because this is not in the CV. You have to talk to the people and to find out in the limited time you have with a candidate. is this there, is it inside the person, kind of

TIM: And once it actually comes to learning and adapting one one of my favorite books, which I can just see on my bookshelf over there is atomic habits. I'm not sure if you've read that one before.

OLAF: I, I heard of book, but don't have the chance to read it, so I think I have to put it on my list.

TIM: It's definitely worth a read. And it's all basically about changing your behavior to stop doing bad habits and start doing good habits. It's pretty simple. And in a way, then, if we're saying that AI for a lot of things is now making us a lot more efficient. And some people are maybe stuck in their ways a little bit and persist on doing things in a way that's now a bit outdated. That's just a habit change of just like, how can I, instead, I used to write manually code every day. From now on, I'm going to open up Claude. I'm going to prompt it to write me the code and there's just a habit change that needs to take place, which if you've been doing something the same way for 30 years, I can understand how it would be challenging to, to make that behavioral change. But for anyone listening, highly recommend that book to change any habit of your life.

OLAF: Yeah. Sounds nice. Sounds nice. But it is it's exactly about this. It is when you look to the person itself to the human, you need this kind of basic habits and this behavior of I'm able to adjust. I always also bring this analogy of of water, like a little bit connected to Bruce Lee, right? You have to be water, my friend, because the water wants to reach the ocean always. but when there's stone in the way, the water will find a way. And this is something what I try to bring with a little bit this kind of picture to my people and say, guys, yes, we follow this principles. We want to contribute with data. We want to unchain the power of data analytics. And the data we have in the organization, but we have to adapt to the situation we have in the organization. And that means when there's something on my way, I have to find a way around, but my goal is still to reach the ocean. Exactly this.

TIM: That's a great saying and analogy, which I hadn't actually heard. So it feels like a startup is like that as well. You just got to find a way, no matter what, somehow you've got to, got to get to that ocean. So. I'm going to add that one to my, my little toolkit of saying, so thank you for sharing that. One final question today I'll ask for you is if you could ask our next guest one question about hiring, what question would that be?

OLAF: Yeah. And in the end, I'm would be very interested to see how other data leaders doing their work. in, in, in this very amazing times we're living in with AI and data. And it's about what the innovative strategies or technologies, the person implemented to in the hiring process process to change significantly the quality of hires, because I think everybody getting so much rules and applications. Because everybody can be now a data scientist to have by, by being in a course from Udemy or something for six months and not have I would say academic foundation behind this. But when we look to this, And implement strategy, a strategy to do the right hiring in a way, how, how much is a change in the improvement of implementing this kind of technology, not just going on to say, Hey, I need a data scientist. It is about what is improvement behind here? And what is the innovative strategy? The person is putting in place.

TIM: Wonderful. Well, I will level that at our next guest. I'm not sure who it is, but it's someone tomorrow. So we'll see what they. What they say, and I'll share their answer with you. Olaf, it's been a great conversation today. We've gone across a lot of different areas. It's been really interesting and thank you so much for sharing with our audience, all of your insights and wisdom.

OLAF: Thanks for having me, Tim, and looking forward to 2025. What will bring this to our area?