115: Rethinking AI Governance for Enterprise Adoption with Dr. Markus Schmidberger

Speaker A: We don't need a chief AI officer. It's a wrong decision to hire one.

Speaker B: Do you think that we're heading in a similar direction for AI adoption and usage in commerce and businesses?

Speaker A: There is a gap between what we promise around AI and where we create business value out of it.

Speaker B: Why do you think that businesses are having a hard time translating AI application to specific use cases, but also then being able to say, no question, this is a win for the organization.

Speaker A: It's not a technology topic, it's a cultural enablement topic. I'm really wondering if a low AI adoption at the moment creates more sustainable companies in the next two to three years than the ones who are opening up AI to everyone and do every crazy thing with AI.

Speaker B: Uh, interesting.

Speaker C: Dr. Markus Schmidt Burger is a data and AI leader, helping companies create real business value from emerging technology while challenging conventional thinking on AI adoption and building Junto AI, a next generation professional network. Welcome to Using AI at Work. I'm your host, Chris Daigle.

Speaker B: Each week we'll be learning how today's

Speaker C: business owners, entrepreneurs and ambitious professionals are getting more done with smart use of tomorrow's tech.

Speaker B: Let's get started.

Speaker C: Right now, every business leader is asking the same question. What are we going to do about AI? If this is you, chiefaiofficer.com has the answer. We give you a simple path forward where we provide executive and team training so your people know exactly how to safely use generative AI in their day to day. We also manage the deployment and implementation to make sure tools actually get adopted and deliver results. And we'll also guide company wide transformation so AI becomes part of your operating system, not just another shiny object. The companies that act now will increase productivity, cut costs and grow faster than their competitors. Those that wait will get left behind. So if you want to make AI work in your business, visit chiefaiofficer.com and see how we're helping companies of all sizes finally get results from AI.

Speaker B: Greetings everybody and welcome back to another episode of Using AI at Work. My name is Chris Daigle and I'm the host. Uh, today we've got an interesting episode. Um, our guest is Dr. Marcus Schmidberger. Uh, he is the founder and CTO. And Marcus, it's Junto or Junto AI.

Speaker A: So Junto is the Spanish word. Junto AI is the one that the English people prefer to say. So we tend to junto AI.

Speaker B: Okay, awesome. So, um, and junto AI is not the reason that we're here today, but it's a startup building what he calls the next generation but business network. Um, Marcus is certainly qualified to talk about the topic today, which I'll reveal in just a moment. Um, with 15 years in data leadership, including stints at AWS Scout 24 where he led a 40 person data organization across pretty robust, uh, data engineering, data science, uh, data access, ML engineering environment. The reason that Marcus is a guest on the uh, podcast today is because I saw a post on LinkedIn where he had some um, very specific positions on the role of the Chief AI Officer. And if you're not familiar with what I do outside of the podcast is I actually founded a company called chiefaiofficer.com and I'm not going to reveal the spoiler, Marcus. So why don't you kind of share um, what your position was in that LinkedIn post. And welcome by the way.

Speaker A: Yeah, welcome and nice to all the listeners here. I'm Marcus and yeah, I'm next to being a CTO and an advisor for data, um, organizations and technology leaders. I like to post on LinkedIn three to four times per week and I have a very opinionated position there. And my post, which was about, I think it was a month, three weeks ago, was about that the Chief AI Officer is a wrong decision to hire one. We don't need a chief AI officer. Uh, that was my post. Um, it was very provocative but it also got 100,000 impressions. So a lot of people were looking at that post. Um, I had about 120 comments. So the discussion was also very intense and it took me busy for, to be honest, for nearly a week. And um, and yeah, you know, that's nice. Part of being showing up on LinkedIn, having opinions that you then talk to people, conversations come up and we ended up deciding to do a podcast about that. So thanks for the invitation.

Speaker B: Of course. So, you know, obviously, uh, being the founder of chiefaiofficer.com and being a big proponent and for those listeners, our definition, it may not be Marcus's definition and we'll find that out, but our definition is, it's a non technical role in my perspective. It's um, somebody who understands business operations and understands the generative AI capabilities, landscape developments and is able to translate the two. Now my first uh, instance of finding the term Chief AI officer was from a Harvard Business Review article, I think maybe in 2015. And at that time there wasn't a generative AI path for AI. It was all like, it was a data science, it was a machine learning, it was a, a technical role. Um, so I Think the first distinction here, Marcus, is that let me clarify what you define as a chief AI officer.

Speaker A: Good, good question. So, uh, where we definitely align is for me, the chief AI officer is a person which is not technical and is able to translate business topics into AI to help an organization to identify where are AI use cases and what we have to do to create business value out of that. I think that's the part where we align very well. And to give you some background, where I'm coming from is the last 20 years I was working for European companies, helping them to create business value out of data, and there is this data business gap. Uh, Gartner was pushing that through all the keynotes over the last 15 years. It was a big topic and we introduced the chief Data officer in a lot of companies to close the data business gap. And um, when I'm now today looking on what's happening with AI and when I talk about a chief AI officer, I think it's a person who is sitting between technology, data and business and bringing them together.

Speaker B: Okay, now you mentioned this data business gap and you're suggesting that there is an AI business gap that's coming, that the next big enterprise problem will be that, similar to that, that long running data business gap where companies built data teams, data warehouses and data literacy programs, but many of them still struggled to make data a part of their daily business. Do you think that we're heading in a similar direction for AI adoption and usage in commerce and businesses?

Speaker A: Yes. So I have written a LinkedIn post about that last week and I predict that for the next five years the AI business gap will be the keynote of Gartner at every conference because we are running in the same direction, um, that AI is driven by a small group within the organization and very often unconnected to the business value. And when we look on the AI use cases today, and I especially have a perspective on the European market, sure the quality of chatbots definitely improved, but that's not AI. Uh, we're using AI a lot for coding, but there we see already the first chief financial officers complaining about the amount of tokens we are burning without creating return of investment. And we look into the real use cases where gen AI or AI agents are used for, for products. So if we talk about software as a service, this is very, very low. And when we look on the big players, they are promising that AI will change all of that. But the use cases are so low, which we see in production. So there is a gap between what we promise around AI and where we create business value out of it.

Speaker B: So I want to address that because you know, uh, we talk to business owners regularly and the three main impediments that we find pretty consistently when we talk to them about why they haven't done more with AI is one, I'd love to do more with AI, but I don't know exactly where to use it. And you just reference. The other two are I'd love to use it, but I don't understand the risk. And the third one is I'd love to use it, but we don't have anybody, we don't have a chief AI officer to help us like uh, uh, sort all of this out. So let's talk about that first one. With the use cases you're. And I agree there are, uh, maybe with the people that we work with, it's small and it's anecdotal but, but generally the conversation is that they're seeing results, but it's not necessarily the results that we can tangibly say, wow, this has transformed the business, the economics of the business, the performance of the business and those sorts of things. Why do you think that businesses are having a hard time translating AI application to specific use cases, but also then being able to say like this is, this is no question, this is a win for the organization.

Speaker A: I see a huge similarity to 15 years ago when we had BI tools bringing into the company where we had a central BI tool, uh, told people use that to understand the data better, to understand your marketing flow, your marketing analysis better. The promise was big. The people in the product organizations, the ones who creating new products who are driving the business revenue, they don't have the time to look into that, to learn about that, to develop their own skills, to use it. So developing the skills to be data driven, uh, there was never enough time in the companies and that's also the reason why, why a lot of companies are still operating data as a very siloed central team where you have the experts and they do the work for all the others around. So the number of really data driven companies where a product manager can use the central BI tool to analyze the data and to understand the data, I assume it's less than 20% of all the companies around the world. And the same, exactly the same is happening now with AI. We have some experts in AI, people who are really fascinated in AI, um, very often coming from the data sector, from the data science sector. But when we then look onto the product management side, the business people, the financial people, they hear about it, but they don't have the time to invest into it, to learn a bit about it, that they can then use it, adapt it, and to brainstorm ideas, to bring it into the product and by that create revenue. So it's again, for me, it's not a technology topic, it's a cultural enablement topic. How we enable the culture within the company to be data driven and AI, I call it AI enabled at the moment.

Speaker B: Okay, there's a difference here because this AI access environment is kind of the opposite of the data warehouse access. Right. You drew an interesting comparison in the previous conversation we had where in the data environment those data warehouses were tightly controlled with really only a few, uh, you know, makes sense with only a few people being allowed access to that. But now AI is moving in the opposite direction where we'll go into companies and they'll say we got everybody a license, like without the training, without any of that sort of thing, um, and without any type of like really strong strategy. Now this kind of addresses what you were talking about just a moment ago with that cultural change and it's creating this tension, let's say, between adoption and being able to control it. So why is this understanding, this paradigm, how it differs from the data warehouse access? Why would you think that that would be important for companies to look at it differently than they did with the data conversation?

Speaker A: Okay, so there are two things for me in there. One is the governance part.

Speaker B: Yeah.

Speaker A: We close down the data warehouses to control the data. Now, as you described it very well, we open up AI to everyone. We will have a massive governance issue soon, how to control all of that. And at the same time, by opening up all of that, everyone has access to any kind of chat. If it's Now Gemini or ChatGPT or Claude, everyone has access to any one of that or even to more of them. But to be honest, to use a chat as gen AI, for me, that's not AI.

Speaker B: Yeah.

Speaker A: Uh, AI starts when I use it as an agent for coding or when we have an agent which is doing something. Um, and AI is not for creating LinkedIn posts or for therapy. Ah, I was reading a study lastly that the highest use case of AI is therapists. People are using it for 83% of their therapist.

Speaker B: Wow.

Speaker A: It. It's nice. Yeah. Ah, but that is for a business perspective, that is not AI.

Speaker B: Sure.

Speaker A: And, um, this is where we have to, in my perspective, is where we have to start educating our employees, our organizations, where is the potential of AI that they learn about it and that they then can Use it. And this probably is also went into my post about, um, the chief AI officer is the wrong one is for me, education is owned by the HR or people department. So HR and the people department has to take a big part of ownership of the AI topic because it's enablement to educate people to train them. And it's not about training them today because tomorrow we will have the next new technology. It's really about having a process in place which keeps our employees regular up to date. It's evangelization and ongoing education.

Speaker B: So let me address that, that HR thing. Totally, uh, agree. In the larger organizations, they're in charge for the development of the employees when it comes to upskilling and things like that. And in your experience or what you're hearing in the marketplace, the HR or the people departments, they're not necessarily driving the. I mean they would essentially be responding to a strategic initiative developed further up the food. Food chain, I would imagine. Is that how it's playing out?

Speaker A: Um, so again, my perspective is on European companies. The one where I see which are able to bring agent AI agents into production, create business value out of that is the ones where the HR departments are actively pushing for AI. They are driving the topic. And that's not enough. Um, if you just have one, that's not enough. And that's also my point on the chief AI officer. There's this high risk. If I hire a chief AI officer, everyone believes, okay, that person is owning AI. This person is going to do AI for us. But that will not work. It has to do done by everyone. The structures, the processes in the company. And when we talked some days ago, we talked about the differences between Europe and the US In Europe we see organization structures mostly very hierarchically. Uh, and if I put on the top a chief AI officer, everyone assumes this person plus his small organization, they are owning AI. Um, which will not work because then we have it centralized, it will be a bottleneck and we will not get it into all organizations. So we have to build something, a way to get it broad over the complete organization structure. And this can be done by a chief AI officer if he gets the mandate from the management to bring it into all organizations. But it can't not be done. If I put it into a central role and I think this person is going to do the magic.

Speaker B: Okay, this makes sense. So this is. Now it's really getting clear for me that why this is a much different issue. Um, because in the data scenario, I'm not a data scientist, I'm Not a technician. So therefore if I needed that business intelligence, I would go to the department that managed that. But in the AI environment where I can get a $20 or $25 a month license for the entire team, having them say, oh, I need to go see the chief AI officer before I can craft uh, a prompt or use it as a thought partner. As some of our guests have talked about, that completely different technology, completely different um, paradigm of application of these tools. How are you suggesting, at least again in the European market where most of your um, uh, experience has been with this topic, how are you suggesting that those companies in that hierarchical, stricter, hierarchical structure introduce um, it. Is it a phased rollout by tier? Is it a company wide initiative with staff and executive team all going through something?

Speaker A: So I do hear the, again the comparison to data. We have this framework of data mesh where we had distributed ownership and we had a data platform team providing the technology as a platform. And then we pushed the ownership of data consumer and producers into the different units of the organization. How we did that, or the companies who did it successfully, they put data analysts and data engineers in a kind of in residence model, or today you would call it field deployment engineers. Uh, you put it into the organization unit, you train the unit, you work with them that they learn how to do it and then that person moves out again because the unit can do it by themselves. If we look into the AI world and I think we have to do it very similar here, and that's the reason why we see that companies like Google and Trofik are pushing for these field deployment engineers to bring people who know how to do it from a business and a technology site into the business units, educate them, train them, and then let them do that. If we talk about big organizations where governance platforms are a big topic, we need a central unit who keeps that thing under control and the rest has to be done by, I call it, in residence enablement models. And I assume this will be the most efficient ones.

Speaker B: Uh, okay.

Speaker A: And for me, when we're looking onto that, this is a process where you put an engineer into each of the units and it can be done in several weeks. So it's nothing which is going for the next three, four years. It's just enabling them on AI to get them going and then they will learn by themselves.

Speaker B: So I like that a lot. One of the challenges that I would see if I was a business owner is I'd say I don't have enough of those, those fd, those in residence talent to go around the company. So therefore I need to, there needs to be sequential, um, this, this, this individual, the small team would need to work with a department, get them tuned up an AI and then move on to another. Or are you suggesting that it's a little bit in each of them and then we go back to the first department and then we do a little bit more, go through the departments and then return. Ah. Uh, do you have a position on what would be the best, most effective approach?

Speaker A: So I, um, when I was working at Scout 24 we had an AI evangelist. Uh, she was a very talented product manager and she also had a small team and before an FTE came into the unit, she went into that unit and they did small proof of concepts which was, she went in for a week or two weeks, a short time to evaluate what are the business use cases for AI in there and then she moved on with her team into the next unit. And once per every six or eight weeks, um, we call it AI board meeting, where we were sitting together with some decision makers from the company and decided which are the proof of concepts, which are the most easy one to implement, which create the most business value. And then we have chosen mostly M2 of them and then put FTEs onto that. And then, so it was a, uh, two level approach. The first level was really that discovery and then it went into the implementation. Huh. And by that in the discovery we educated the people about what it is about in the implementation. We educated the teams a lot about what are the responsibilities, how do we operate it, how can we really create value out of that. And then we have the third phase where we brought it into production and in the time of the production we kept um, a regular checkpoints with these units to see if it really works and if it really delivers the business value we evaluated in the first phase.

Speaker B: Okay.

Speaker C: Ah.

Speaker A: And by that we move from unit to unit as well.

Speaker B: Yeah, yeah. So it's, I like this approach because that initial person is somebody that, yes, they're the AI enthusiast, but they're not, they're able to speak the same language as staff level employees. They're not technically minded and having trouble translating those technical, uh, considerations and terms into standard, uh, employee language. Okay, I like that a lot. Secondly, by them coming in and not necessarily saying we're going to do it all today by helping those individuals, uh, source those potential pilots or use cases, um, that's very like, that's not threatening, that's, oh, we're looking for places where AI can help you on a day to day. I like that a lot because as a change management consideration, we're easing them into the conversation as compared to say, in like the next two, three weeks is going to be really intense because we're going to identify pilots and we're going to start to build your skill set and we're going to start to build the tools or the solution and that sort of thing. So I like this approach a lot.

Speaker A: Um, and apart, what I loved a lot as head of data at that time was it helped me a lot in the budget process because by evaluating every two weeks, this proof of concept, in the end we had a long list of use cases, we had evaluated all of them, we measured them against business value and I could use them at the end of the quarter or uh, half year meetings to go to the board and tell them, hey, have a look. This is what we have done. We create that business value. This is the opportunity which is sitting there, which you can enable when you give us more budget for more headcount.

Speaker B: So, and this is something that's come up in other conversations and that, that has been how do we measure the ROI of this? Right. Because, um, it seems like it would be pretty easy, but when you've got these individuals. Look, if I'm working on a pilot project that's very discreet, I get it. This is how many units, how much time, how many resources were required to produce the widget before AI. This is what it looks like after, what's the delta, that's our savings. Totally understand that. But when I'm teaching these people, some of the individuals are going to get it. We call it thinking in AI. They start thinking in AI sooner than others and they start saying, oh, if I can do that with this, oh, then maybe I can do this and I can do this. And they're doing it just kind of dynamically throughout the day as things come up. Is that an impact? Sure. Can I measure it? I don't know. Do you have, and this is kind of a little, uh, of a jag from our conversation, but on that topic of ROI measurement, did you see that come up at all? Did the company just say it's too hard to measure? We're not going to necessarily worry about that. We'll just focus on the pilots. Does that question make sense?

Speaker A: The question makes sense. And, um, I think it's a hard question at the moment because at the moment we measure a lot in token usage. Huh. That is the number most companies are aware of. How many tokens do we burn per day or Per week, but they describe nothing about the outcome. It's. And I think it's because another one of the. In this list of the mostly used use cases were improving. My email, um, was number four of five in the most use cases for Chennai. Um, yeah, it's nice to put your email into the gen and ask the AI to write it in a better tone or everything, but this is not creating business value. Um, what, what I see with, with companies who are able to create business value out of that, they, um, they don't talk about tokens anymore. They talk really about agents. Because when we deploy agents, we then have a kind of mechanism which is doing work for me.

Speaker B: Okay.

Speaker A: Uh, and, and which is more than just writing an email better. And they measure the impact of the agents. So when an agent makes your. One example I have is the. I have one company which improved the, the travel experience for their employees. So before they had to go to a portal, search for the hotel, uh, search for the flight. Now they enable the booking agent which looks into the calendar and directly proposes you the flight and the hotel. And you just tell the agent, yes, do that and they don't have to do anything else. And this saves the employee the time. And it's a clear agent which is doing work. And by that we can measure the impact of the agent. And we are not talking about tokens anymore.

Speaker B: Okay, now this makes sense. Um, because that is the conversation that we're. I'm starting to see a lot in, you know, companies saying we spent our entire Token allotment in 30 days or, you know, we spent a quarter's worth in a couple of weeks. And yeah, okay, great. That indicates people are using the, the tools for sure. But what are we getting out of that? I like this idea of rather than looking, using token spent, certainly like you have a budget or whatever, but not necessarily trying to equate the token spend, but looking more at the agent outcome is much, uh, cleaner because you can equate token. I can't equate that to an employee. That's, that's like a budget. But an agent, if they anthropomorphize it, is like a co worker. Right. It's like somebody doing something. So it becomes much easier for me to translate how long would it have taken the human to do that versus the AI employee of the agent? Okay, there's.

Speaker A: I, I have to look after it. I was reading a blog post some weeks ago that there is a correlation between educating your organization on AI and token usage. So the more you have educated your organization on AI how it works it the more it reduces the amount of tokens which get burnt. Ah, uh, yeah, it was a blog post about strongly technology companies which use genai or AI a lot for coding. And um, in coding you can improve the token usage a lot by knowing which task or what requires how many tokens. So by educating people on that you really can reduce the token usage. I assume this holds for all areas?

Speaker B: Yeah, now that makes perfect sense. And it's a further endorsement for the, either the hr, the people team or some individuals internally to push for uh, upskilling AI enablement, AI literacy fluency for the teams so that not only are they able to use it better but they're actually using using it more, they're more fiscally responsible with token application. Okay, now you know one of the things that, that some of the listeners may be saying was well we've got chat GBT licenses, we've got CLAUDE licenses. Are we going to run out of tokens? And so I think what might be an interesting um, uh, definition or explanation right now would be the difference of environments where generative AI application is included in my monthly subscription versus those pesky uh, tokens being spent. Can you take a second and maybe explain a couple of scenarios where hey, if you're doing this you're safe, you don't have to worry about getting some big uh, big bill from OpenAI for your tokens. But if you're doing this, this is where you need to also be token aware.

Speaker A: So most companies I see were where the CFO is complaining about the token usage are technology companies where AI is suddenly used for all the coding. So this is definitely, it's 80% of the cases. I would prefer to put it in a different direction. Um, the token usage, it's a pain at the moment. Fully agree but there are also that millions of companies where the employees are not allowed to use AI um due to concerns of the management, due to government's reason and so on. And for me this is at the moment this, it's like this big AB test. From my perspective we're doing the biggest worldwide a B test on a ah, company level companies who are using AI. And when we listening to anthropic Google and all of them they will have a massive advantage in the next 12 to 18 months. But at the same time we have all these companies who don't have access to AI due to regulations, due to beliefs by the CEO and so on. And again everyone says these companies will die. I'm not, uh, six months ago, I would have told you the same. Today I'm not sure if this is going to happen because with AI people are getting very often off track. Uh, these companies are losing the traction to their strategy, which are using AI heavily. Uh, and they might open up a new market, but might also not be focused on their core market anymore. By the way, the companies who are not using AI, they keep on working in the same way. They have the same processes, they know how, they have built revenue over the last 10 years and they keep on going. I'm really wondering if low AI adoption at the moment creates more sustainable companies in the next two to three years than the ones who are opening up AI to everyone and do every crazy thing with AI.

Speaker B: I have not heard that position, but as you explained it, so let me just make sure I'm m understanding this. There's this risk if we give everybody AI access that it becomes a distraction from their core function inside the hierarchy, inside the organization. Yes, interesting.

Speaker A: Um, and when, when I talk, I talk to, I talk to. Over the last four, four months or ah, five months already, I talked to several product leaders and companies and they all told me how great AI is, how easily they can build a new proof of concept, they can create a new ux. But when I talk to them closely, I all learned from them. They explored new market segments but didn't focus on improving their existing business.

Speaker B: Wow, uh, wow.

Speaker A: And this is for me, I'm getting scared about that, what's happening there, even when I look into the startup world. So I'm at the moment as a CTO of a startup, I'm very much in that startup world where everyone spins up within a weekend a new product and brings it on the market. And there is um, from the Google Play Store, there was a statistic which was released some weeks ago. The numbers of apps on the Google Play Store exploded. The satisfaction of the quality of the apps dropped.

Speaker B: Yep, I can see that.

Speaker A: And this is exactly the impact of AI. And if that is happening with all the businesses where we give them full access to AI, um, they explore more niches, but the quality will drop.

Speaker B: Huh.

Speaker A: And what builds up trust and what brings customers. Quality.

Speaker B: Yeah, yeah, interesting. I hadn't considered that, but I totally like, I think that's a fair assumption because I see that um, like we've got some clients that we've been working with for a while now and their teams have really gotten capable with the tools. They went from, you know, basic, like what is a prompt to now they're saying, oh, over the weekend I got on AI Studio or I got on cloud code and I built this thing. And the automatic assumption, I think, from everybody is that there was consideration and judgment that went into, well, what should I do with the tools? Ah, uh, the demand or the need is this, let me build it. But instead it could be, ooh, I can build this. Well, what's the business impact? I. I don't know. But that doesn't change the fact that I can build this, so let me build it. Huh? Do you have any, um, when you're hearing this from the people that you're talking to, are they aware that of what they're telling you when they say that, are they, are they seeing the same perspective that you are?

Speaker A: They are all wondering where their business model and company structure is going in 2027. Because most, most companies operate like that. I define in the beginning of the year the strategy for 2026. I execute against that and I do a new strategy in 2027. And to all of them I talked to were like, it's really interesting to see what kind of strategy we do in 2027, if we really keep that broad or if we just go more back to our core and focus back on the core. That was the constant message I got from everyone. Um, the strategy for 2027 will be really interesting and, um, I personally think that there will be some companies who will drastically go back and skip access to AI to focus back on the core product.

Speaker B: So I'm processing this right now. And as somebody who's working with companies and going in and helping develop strategy and then working, you know, from the bottom up on where does this, where is it actually painful and where does this use case or this particular pilot map to strategy, um, this is, this perspective that you just introduced is causing me to, uh, do a temperature check on the conversations that we're having with newer users and helping them understand it's not about the possibility of like, now it's been expanded, it's been, let's apply this possibility specifically to the things that you're already doing as compared to what's possible. Um, so listeners, if you've been looking for struggling for use cases with AI, it may be because you're thinking about the possibilities as compared to the specific applications that are already in front of you, the things that are known issues, friction points, constraints within the organization and process. And instead of looking for the new things, just take the AI and start saying, how do we use generative AI in this process? That we're already doing. Now. One thing that I would say as a caveat is don't assume that the process that is happening now is the most efficient process. So don't necessarily just say how do we throw AI into a dysfunctional process? Evaluate the process first, make sure it's optimized and then introduce. So, um, I like that a lot. So uh, Marcus, let's talk about Junto now. It's kind of a bit of a pivot because the, the startup that you're involved in as a CTO isn't necessarily, I mean I guess you could say it is kind of a data play behind the scenes. But the user benefit isn't, has nothing to do with data, it has to do with uh, connections with a more effective environment for connecting with other professionals.

Speaker A: So what we have done is we looked onto how people are doing business networking today, how Chris and Markus connect, what kind of messages they send to each other, how they built up trust, how they find each other. This by today when we talk to recruiters, salespeople. This is a manual process by writing emails, by searching people on LinkedIn which also a lot of random in there to bet. Yeah, Chris is a match to me. I sent him a message, I don't know you anything. Uh, and um, it takes some time if you get connected and something new is happening. And we said let's rebuild that model with agentic AI. So we are not building on top of LinkedIn or business networking. We said we are building something completely new with AI agents. So how it works, every human gets a digital twin. So there's a digital twin, Marcus. There's a digital twin, Chris. And then when Marcus for example wants to hire an engineer, my digital twin will talk to your digital twin. If you're an engineer for me, it probably knows directly that you're not an engineer for me. But then your digital twin will search in your network if there is an engineer for me. So by that we get in seconds in the 2nd, 3rd, 4th, 5th level of network. So we are not finding an engineer which is the best match for my network. No, we find an engineer which is the best match around the world for me. And by training the digital twin, we are not only searching for skills or hard facts. No, we are also searching for values, red flags, um, communication, style. And that works for hiring, that works for sales and everything. And when we take for example this sales example, salespeople spend a lot of time researching you before they jump on the call the first time so that they know how they can Quickly create connection to you. If they have to talk about the weather or your kids or whatever, the AI agents are doing that for you. And only when your digital twins finding out that there is a match higher than 80%, then they introduce the humans to each other. Because we believe humans buy from humans. Ah, so we are not removing the humans, but all that manual, slow, time consuming process we do before we now do wire AI agents. Um, there's a division of Gentle AI.

Speaker B: Yes. So, um, how soon do you expect this to start having users on the platform? Because I like this concept a lot. It seems much more effective and efficient than let me go to LinkedIn, let me go to Sales Navigator, let me do a search term based on a keyword. Um, if my agent can go out there, not only connect with you, but if you're not a fit, tap into your network on my behalf. That seems much more effective.

Speaker A: Yes, exactly. And this is what we are promising. Ah, where we are at the moment is we have a prototype, uh, we have first users on it. We are looking at the moment for an investor to invest into that vision that we then can build up a team and get it going. Um, if we have an investor, we assume that it will take us around four to six months to make it open for everyone. Because the challenge of building a new business network is it's a two sided marketplace. Uh, you need people offering something, people searching for something and it's not working. When we have 10 people in there, we need around 50,000 people in there. It's called the cold start problem. So we have different approaches in place to solve that. But we have to collect these 50,000 people and as soon as we have them we can open it up globally and then grow and grow and grow.

Speaker B: Well, when you're ready to start taking those on, I'd certainly love to volunteer the chief AI officer community that we've got. It's not 50,000, but it is, um, it's a few thousand people that are eager to be early, early in on new technology, specifically AI, but would probably have the types of networks that others would be looking for. Because I would imagine that with the name Junto AI that individuals will understand, hey, there's an AI element as a user, there's an AI element to this that I need to leverage and quite possibly let me find somebody with AI talent in the network because that's what everybody needs, more people with AI talent. So I would imagine that initially there will be a lot of searches for AI talent in that platform.

Speaker A: Yes, exactly. This is One of the ideas we are following there.

Speaker B: Yeah, very cool. Um, so seriously Marcus, if we can um, be part of that beta group or whatever you need, I'd love to um, introduce that.

Speaker A: I will reach out to you. Definitely.

Speaker B: Community for sure. Um, so this is great Marcus. It's been in such a short period of time. I've had a number of takeaways that have certainly helped me understand the broader landscape of the uh, generative AI conversation. But more specifically, as somebody who's bringing it into companies, an AI implementation lead and that sort of thing, it's given me a ah, different perspective, challenged uh, some biases that I didn't even realize that I was holding about how to do this thing and m. Perhaps uh, an improvement to the approach that we take when we're bringing AI into it. So thank you for, thank you for being so polarizing in your initial post that led to us being on this conversation today. So for those who want to kind of follow what you're doing with Junto, but also pay attention to the commentary that you're putting out on LinkedIn. What's the best way for them to uh, uh, stay on top of what's up with um, the efforts that you're putting into this?

Speaker A: Just search on LinkedIn for Dr. Marcus Schmidt Burger. Click the double bell so that you get all my um, my posts and um, follow me on LinkedIn. This is a way to, to stay connected. What's happening. Um, and I thought I post a post about data and AI, um, enablement in companies and about joint AI. This is my CTO role at the moment. Yeah.

Speaker B: And for the listeners we'll have all of this in the show notes. And my final endorsement would be um, based on the interaction that Marcus is getting with his post obviously indicates that he's addressing some part of the conversation that's happening globally that's not necessarily being represented by others. So I would encourage uh, you to keep um up with what Marcus is doing. Not just with his um, perspectives that he's bringing from the data environment, but also the direction of Junto, which I think is um. Now that I understand it better, I'm particularly eager to be an early user of that. So thank you for sharing with everything and any. I guess for those that are listening to this, because we've got people that are listening, they're at all stages of the AI journey. Any final advice for them?

Speaker A: Uh, yeah, my final advice is open up your AI and ask the AI the prompt, give me a random number between 0 and 100. It will be either 42 or, uh, 73. Why? Because these two numbers are happening most often in the Internet. So AI is a statistical model. Um, and keep that always in mind that the AI is just a statistical model. And, um, you can ask the AI as well why is that the case? And it will explain it very nicely to you. And if you have got that, you will see, you will learn where the power for AI is really sitting. Uh, and that's my advice, which most people really like at the end.

Speaker B: Yeah, so while you're. If you're listening to this in front of a computer, please do that, that prompt right now. It's almost like a magic trick. You'll be surprised at the accuracy of what he just said. And it's supposed to be a random number, right? Well, maybe it's not so random. Well, Dr. Marcus Schmidt Berger, thank you so much for being a guest on this and taking the time out of, um, your busy schedule of, uh, swims, um, and such, enjoying your summer to be able to share this perspective with the listeners of the, uh, Using AI at Work podcast. And again, folks, um, take a look at the show notes. Uh, add Marcus to somebody that you follow and you will certainly, uh, benefit from his perspectives when it comes to better understanding this conversation happening globally about generative AI. Thanks, everybody. We'll see you next week on another fantastic episode of Using AI at Work.

Speaker C: Thanks for tuning in to Using AI at Work. Don't forget to subscribe for more conversations about how to use AI at work. And a special thank you to our sponsor, Chief AI Officer for empowering businesses with AI education and training. Visit their website for a free AI Readiness Assessment and AI Strategy Guide to help you get started using AI at work. That's www.chiefai officer. Follow us on Twitter at the handle usingaiatwork and visit www.usingaiatwork.com for free resources to help you harness AI in your role.