1 00:00:14,120 --> 00:00:17,840 Hello, this is Eva and you're listening to the In Between Tech 2 00:00:17,840 --> 00:00:21,240 and Trust podcast. And this week's episode gets 3 00:00:21,240 --> 00:00:25,320 quite practical because I talked to Doctor Mark Roman Frankel and 4 00:00:25,320 --> 00:00:29,960 he's partner and Associate Director at BCG, responsible for 5 00:00:29,960 --> 00:00:32,640 the AI and digital transformation practice. 6 00:00:32,800 --> 00:00:36,240 And together with him, we dive into one of the latest studies 7 00:00:36,240 --> 00:00:40,120 that was published in September 2025, and it's called The 8 00:00:40,120 --> 00:00:43,280 Widening Value Gap. And if you're like me and you 9 00:00:43,280 --> 00:00:46,880 read a lot about AI and tech transformations, you've heard 10 00:00:46,880 --> 00:00:52,160 about the stat that only 5% of businesses actually gain value 11 00:00:52,160 --> 00:00:57,440 from their AI projects. And Roman was one of the experts 12 00:00:57,440 --> 00:01:02,040 that interviewed good approximately over 1200 13 00:01:02,160 --> 00:01:06,880 organizations and found out that this gap actually exists and 14 00:01:06,880 --> 00:01:11,160 that the statistic is representative for how AI use 15 00:01:11,160 --> 00:01:15,720 cases are being incorporated or how they also fail. 16 00:01:15,880 --> 00:01:19,680 So we dove a lot into what it means when organizations talk 17 00:01:19,680 --> 00:01:24,280 about AI transformation, how they get AI ready, what it means 18 00:01:24,280 --> 00:01:30,520 to put an organization into an AI first status, and how 19 00:01:30,680 --> 00:01:35,600 processes and people are the means to a successful AI 20 00:01:35,600 --> 00:01:41,440 transformation. Welcome Roman, I'm so glad to 21 00:01:41,440 --> 00:01:44,080 have you. Hello, if I thanks for inviting 22 00:01:44,080 --> 00:01:48,120 me, I'm very happy to spend the next couple of minutes or hour 23 00:01:48,120 --> 00:01:51,200 with you. And before we dive into like 24 00:01:51,200 --> 00:01:56,920 your new findings, you've just launched a huge report on AI and 25 00:01:56,920 --> 00:02:00,160 the use cases and also the findings that you have and 26 00:02:00,160 --> 00:02:02,240 guards of the implementation and so on. 27 00:02:02,240 --> 00:02:06,880 I do want to start with you with a quite overarching question, 28 00:02:06,880 --> 00:02:11,240 because what does trust mean to you in the context of AI 29 00:02:11,240 --> 00:02:15,080 transformation? And is there a way or has it 30 00:02:15,080 --> 00:02:18,480 also evolved over time on how you viewed it? 31 00:02:18,840 --> 00:02:22,520 I think that's a pretty profound question for the first as a 32 00:02:22,520 --> 00:02:25,560 beginning and prior some some reflection. 33 00:02:25,680 --> 00:02:29,520 I personally experience trust when priorities are pretty 34 00:02:29,520 --> 00:02:34,320 clear, so and companies or people see that they work 35 00:02:34,320 --> 00:02:38,320 towards value, which is tangible and when individuals are 36 00:02:38,320 --> 00:02:42,120 recognized and rewarded. And when in these transformation 37 00:02:42,120 --> 00:02:44,640 types of work, everything is very much solution and not 38 00:02:44,640 --> 00:02:47,160 probably oriented. So it has a positive spin. 39 00:02:47,520 --> 00:02:52,240 Sometimes I see it as a risk for trust if things have to be too 40 00:02:52,240 --> 00:02:55,280 fast and too bold. Sometimes people like also 41 00:02:55,280 --> 00:02:59,400 reliability and steady progress and experience. 42 00:02:59,520 --> 00:03:03,440 Yeah, resilience in activities and it's it's quite fun. 43 00:03:03,440 --> 00:03:05,720 I mean we both know each other from the House of beautiful 44 00:03:05,720 --> 00:03:09,280 business and we have attended great sessions by the famous 45 00:03:09,280 --> 00:03:14,520 coach Michael Bunyas and he and he did a great research and when 46 00:03:14,720 --> 00:03:18,680 situations, relationships are trustworthy and love lasting. 47 00:03:18,680 --> 00:03:21,720 And I think resilience was a key password that took take away. 48 00:03:21,720 --> 00:03:24,880 So when everything doesn't go right, people in an environment 49 00:03:24,880 --> 00:03:30,040 of trust, they know that things progress and that resilience is 50 00:03:30,040 --> 00:03:33,000 key in this this setting. So when, when, when I look into 51 00:03:33,000 --> 00:03:36,560 what what has changed over the last years, I mean AI 52 00:03:36,560 --> 00:03:38,920 transformation or digital transformation or IT 53 00:03:38,960 --> 00:03:44,520 transformation before has always been organised large programs 54 00:03:44,520 --> 00:03:48,200 and major investments. I think taking lots of risks, 55 00:03:48,200 --> 00:03:52,560 but also promising big change. I see that nowadays things need 56 00:03:52,560 --> 00:03:56,600 to be more Swifty like budgets are tighter, time is more 57 00:03:56,600 --> 00:03:58,720 scarce. So if you want to build trust 58 00:03:58,880 --> 00:04:02,680 with your seniors, you need to prove value very fast. 59 00:04:02,680 --> 00:04:04,680 So you need to demonstrate impact. 60 00:04:04,760 --> 00:04:08,760 So you can't just expect that that people wait for years until 61 00:04:08,760 --> 00:04:10,800 you get an ROI of any transformation. 62 00:04:10,800 --> 00:04:14,520 So I think trust in AI transformations nowadays is 63 00:04:14,520 --> 00:04:18,680 really getting things done as talking more doing and proving 64 00:04:18,680 --> 00:04:22,800 value very fast. And you've just described what 65 00:04:22,880 --> 00:04:26,480 we could call a widening value gap, right? 66 00:04:26,600 --> 00:04:29,320 Because there's on the one hand side, like the speed, on the 67 00:04:29,320 --> 00:04:31,880 other side, like the steadiness that we want to have. 68 00:04:31,880 --> 00:04:35,920 But it's both also in an economic and an ethical term. 69 00:04:36,000 --> 00:04:40,600 And when you like also from your experience, look into today's AI 70 00:04:40,600 --> 00:04:44,800 landscape, what is the single biggest driver of like this 71 00:04:44,800 --> 00:04:49,320 specific gap, particularly in terms of technology leadership 72 00:04:49,400 --> 00:04:52,880 or the mindset that is required? I mean, as you know, I mean I'm 73 00:04:52,880 --> 00:04:59,440 work for BCGI, am driving AI and digital strategy topics and we 74 00:04:59,440 --> 00:05:03,080 have, we are following a core principle which is called the 75 00:05:03,240 --> 00:05:07,280 10:20 seventies principle. We always say that 10% of the 76 00:05:07,280 --> 00:05:09,320 success comes from great algorithms. 77 00:05:09,320 --> 00:05:14,760 Fine, 20% comes from outstanding technology, but 70% is truly 78 00:05:14,760 --> 00:05:18,160 about people, progress and how you organize. 79 00:05:18,280 --> 00:05:22,240 And in the center of the 70s is clearly the sea level alignment 80 00:05:22,240 --> 00:05:24,880 and leadership fluency in AI and digital. 81 00:05:25,000 --> 00:05:28,400 And you need to implement the culture that embraces change. 82 00:05:28,640 --> 00:05:31,840 So success and short start at the very top. 83 00:05:32,080 --> 00:05:35,600 I mean today technology matters more than than ever. 84 00:05:35,720 --> 00:05:40,640 We have seen great companies creating great ideas, thriving 85 00:05:40,640 --> 00:05:44,680 concepts, do rapid innovation, experiment, have great faith, 86 00:05:44,680 --> 00:05:48,840 fast culture and they they build impressive prototypes and 87 00:05:48,840 --> 00:05:51,000 customers love it and everything gets adopted. 88 00:05:51,000 --> 00:05:54,560 But suddenly they don't scale and the question is why? 89 00:05:54,840 --> 00:05:57,360 And often the tech stack was neglected. 90 00:05:57,720 --> 00:06:01,480 So the data is siloed, end to end processes breakdown. 91 00:06:01,520 --> 00:06:03,200 There's lots of menu interventions. 92 00:06:03,200 --> 00:06:06,840 So as I say, everything needs to start from the top and from the 93 00:06:06,840 --> 00:06:08,880 business. You need to protest by value, 94 00:06:09,040 --> 00:06:12,920 but you also need to make sure that the foundation is, is, is 95 00:06:12,920 --> 00:06:14,680 right. So the tech and data. 96 00:06:14,920 --> 00:06:18,320 So if you have both, I think then then you make it or at 97 00:06:18,320 --> 00:06:22,440 least the companies who don't have both, they see their scale 98 00:06:22,520 --> 00:06:26,360 inhibited after a while. And you've just touched upon 99 00:06:26,360 --> 00:06:30,800 some of the findings that you've like collaborated on with like 100 00:06:30,800 --> 00:06:33,840 the report or also your clients and the role that you have at 101 00:06:33,840 --> 00:06:36,240 BCG. But also there is like quite a 102 00:06:36,240 --> 00:06:41,560 specific number that most of like us have heard about the 5% 103 00:06:41,560 --> 00:06:46,720 of companies that generates a stem substantial value from AI 104 00:06:46,760 --> 00:06:52,400 and like the rest actually is doomed or is set to stagnate. 105 00:06:52,440 --> 00:06:56,640 And can you share with give us a bit more background of what 106 00:06:56,640 --> 00:07:02,280 differentiates those 5% of organizations also in terms of 107 00:07:02,280 --> 00:07:05,480 looking beyond just the resources, I was wondering if 108 00:07:05,480 --> 00:07:10,800 it's like more about structure, courage or potentially also the 109 00:07:10,800 --> 00:07:14,640 clarity of ambition towards their road map ahead? 110 00:07:14,920 --> 00:07:18,720 Yeah, great question. And then to maybe take one step 111 00:07:18,720 --> 00:07:22,560 back and give the audience some context, for a decade around 112 00:07:22,560 --> 00:07:26,600 now, we are serving every year senior executive, senior leaders 113 00:07:26,720 --> 00:07:30,720 covering all industries, all regions and serving them where 114 00:07:30,720 --> 00:07:33,840 they stand in digital. And of course, nowadays with AI, 115 00:07:33,920 --> 00:07:37,360 we check how they invest, we check how they succeed, we check 116 00:07:37,360 --> 00:07:39,800 how they measure value, but also how they leverage tech and 117 00:07:39,800 --> 00:07:43,040 technology and data. And in this year's study, which 118 00:07:43,040 --> 00:07:48,240 is actually called dividing AI value Gap, we surveyed 1250 119 00:07:48,240 --> 00:07:52,640 companies and you mentioned only 5% of them are really realizing 120 00:07:52,640 --> 00:07:54,720 meaningful business value from AI. 121 00:07:54,720 --> 00:08:00,040 So this is like I think it was 68 companies out of the 1250. 122 00:08:00,120 --> 00:08:04,160 Of course, they invest more into AI, but the real differentiator 123 00:08:04,160 --> 00:08:07,920 is that they strategize and act differently and they don't read 124 00:08:07,920 --> 00:08:12,840 AI just as a technology project, but as a longer full business 125 00:08:12,840 --> 00:08:15,640 transformation. They did operating models where 126 00:08:15,640 --> 00:08:18,920 the data, algorithms and people work together seamlessly 127 00:08:18,920 --> 00:08:21,440 integrated and they set the right ambitions. 128 00:08:21,520 --> 00:08:25,720 They focus on those areas where I creates the most significant 129 00:08:25,720 --> 00:08:28,040 advantage. And they also don't just focus 130 00:08:28,040 --> 00:08:31,120 on small incremental productivity gains, like making 131 00:08:31,120 --> 00:08:34,480 a big office function more efficient, but they really focus 132 00:08:34,480 --> 00:08:37,480 on reshaping the core of the business. 133 00:08:37,480 --> 00:08:40,559 That's a bigger task to crack, but much more valuably at the 134 00:08:40,559 --> 00:08:43,080 end. And as I heard your question 135 00:08:43,080 --> 00:08:46,480 about structure, courage and clarity, but I actually observe 136 00:08:46,480 --> 00:08:50,520 is the clarity of ambition is by far the most specific factor. 137 00:08:50,560 --> 00:08:52,240 I mean, everything follows from it. 138 00:08:52,320 --> 00:08:56,360 I would even say that courage is you need to be gracious to also 139 00:08:56,360 --> 00:08:59,120 have clarity because you, you can't just do minor small 140 00:08:59,120 --> 00:09:02,880 investments or have some chase, some single digit efficiency 141 00:09:02,880 --> 00:09:05,200 gains. You really need to be clear 142 00:09:05,200 --> 00:09:07,840 where you put your, your place, what are your board 143 00:09:07,840 --> 00:09:11,080 transformative moves and, and where you want to have double 144 00:09:11,080 --> 00:09:13,360 digit growth and faster innovation. 145 00:09:13,760 --> 00:09:17,520 So I think courage, clarity go goes, goes hand in hand. 146 00:09:17,520 --> 00:09:20,360 And of course, if you have that, you need to build the right 147 00:09:20,360 --> 00:09:23,360 structures to execute. And that's why companies 148 00:09:23,360 --> 00:09:25,840 eventually need to build an AI first organization. 149 00:09:26,320 --> 00:09:30,000 AI first is like a term that we've have heard many times in 150 00:09:30,000 --> 00:09:34,480 the past, but there are also those ones who claim to be AI 151 00:09:34,480 --> 00:09:36,920 ready. And I do think that putting AI 152 00:09:36,920 --> 00:09:39,440 first doesn't mean that you're AI ready, right? 153 00:09:39,440 --> 00:09:44,440 So what's the key shift required to move somewhat from, let's say 154 00:09:44,440 --> 00:09:48,720 an experiment or mentioned kind of approach to an enterprise 155 00:09:48,800 --> 00:09:50,880 wide transformation? Good point. 156 00:09:51,280 --> 00:09:54,480 You are right. AI ready first means, I mean, 157 00:09:54,480 --> 00:09:58,000 it's good when you're AI ready. So you already start to run 158 00:09:58,000 --> 00:10:00,960 pilots. You, you, you upscale talent, 159 00:10:01,240 --> 00:10:03,280 you maybe start to build a data platform. 160 00:10:03,880 --> 00:10:07,920 But the pivot to shift to be AI first means that eventually your 161 00:10:07,920 --> 00:10:11,440 entire business model depends on AI to operate. 162 00:10:11,440 --> 00:10:14,960 So AI is not something you used to optimize, but it's really 163 00:10:14,960 --> 00:10:17,280 core. And everything you do has an AI 164 00:10:17,280 --> 00:10:19,200 first paradigm when you design it. 165 00:10:19,520 --> 00:10:22,760 It's not about asking all where we can apply AI and our business 166 00:10:22,760 --> 00:10:26,400 model to become a bit better, have incremental benefits, but 167 00:10:26,440 --> 00:10:29,080 you should really think about how should our workflows, 168 00:10:29,080 --> 00:10:32,440 decisions, the customer experience, our entire company 169 00:10:32,520 --> 00:10:36,720 like be redesigned to succeed in an high power word that requires 170 00:10:36,720 --> 00:10:38,960 fundamentally shifting how you work. 171 00:10:38,960 --> 00:10:42,960 And it means also embedding AI into every process than just 172 00:10:43,000 --> 00:10:45,560 laying it on top of it. And that's why you also 173 00:10:45,560 --> 00:10:49,720 establish eventually trust in AI because grounding it in very 174 00:10:49,720 --> 00:10:53,760 robust data, you implement these reliable algorithms and the 175 00:10:53,760 --> 00:10:55,840 humans and AI seamlessly collaborate. 176 00:10:55,960 --> 00:10:59,160 And then if when they are people eventually and employees and 177 00:10:59,160 --> 00:11:02,640 also customers and everyone sees AI as a colleague and not just a 178 00:11:02,640 --> 00:11:06,880 tool, then adoption accelerates. And what I learned from our 179 00:11:06,880 --> 00:11:10,640 recent study is if we think about what really makes AI 180 00:11:10,640 --> 00:11:14,600 organizations first, AI first is they have much more governance 181 00:11:14,600 --> 00:11:18,320 in place to measure AI value. So, but because they can prove 182 00:11:18,320 --> 00:11:21,280 its value, they dedicate significant more time in 183 00:11:21,280 --> 00:11:24,120 learning and, and, and, and protecting also time for 184 00:11:24,120 --> 00:11:27,640 upscaling employees. They have, I think the number 185 00:11:27,640 --> 00:11:31,360 was three times more centrally defined data policies. 186 00:11:31,360 --> 00:11:35,400 So they have data offices in place and eventually that helps 187 00:11:35,400 --> 00:11:38,920 them to to scale five times more workflows and deploy them. 188 00:11:39,040 --> 00:11:42,200 And the biggest, the by far biggest number was the 189 00:11:42,200 --> 00:11:44,680 engagement of the C-Suite. And it brings us back to our 190 00:11:44,680 --> 00:11:46,800 first question that everything starts from the top. 191 00:11:46,840 --> 00:11:50,120 I think among the share of future build companies which we 192 00:11:50,120 --> 00:11:53,400 use these or name these companies which are AI first, 193 00:11:53,440 --> 00:11:56,800 over 10 times more of the sea levels in these organizations 194 00:11:56,800 --> 00:11:59,280 are engaged versus companies who are lagging behind. 195 00:11:59,280 --> 00:12:02,480 So it's again comes back to the point it needs to start from 196 00:12:02,480 --> 00:12:05,720 from the very top. But at the same time, your study 197 00:12:05,720 --> 00:12:09,680 also points out one of the buzzwords of like, let's say, 198 00:12:09,680 --> 00:12:13,600 2025, and that is agentic or agentic services. 199 00:12:13,600 --> 00:12:18,000 And it's claimed to be one of the biggest accelerators of also 200 00:12:18,000 --> 00:12:20,600 the value gap that organizations can foster. 201 00:12:20,600 --> 00:12:26,120 And in your opinion, what makes agentic AI different and how can 202 00:12:26,120 --> 00:12:30,400 companies use it without falling into let's say complexity traps? 203 00:12:30,400 --> 00:12:34,560 Also like the paralyzes that governance structure might bring 204 00:12:34,560 --> 00:12:37,760 with it and potentially also translate it into more of an 205 00:12:37,760 --> 00:12:42,040 intelligence choice architecture rather than make a process 206 00:12:42,040 --> 00:12:44,640 driven approach. I mean, first, I mean, what is a 207 00:12:44,640 --> 00:12:48,400 genetic AII mean? It is the systems that can now 208 00:12:48,440 --> 00:12:52,120 act autonomously and have still clear goals defined. 209 00:12:52,160 --> 00:12:56,720 And it, it completely reshapes how I functions because it's, 210 00:12:56,720 --> 00:13:00,080 it's not just longer about automation or having the new 211 00:13:00,480 --> 00:13:04,640 creative at it's about not AI that can really act and decide 212 00:13:04,640 --> 00:13:07,320 and execute everything across different systems. 213 00:13:07,320 --> 00:13:10,680 We have this analogy with the brain, the human brain, where 214 00:13:10,760 --> 00:13:12,920 predictive AI is really the left brain. 215 00:13:12,920 --> 00:13:15,560 It's logical rule based. It's very structured and 216 00:13:15,560 --> 00:13:19,280 generative AI is, is the crazy one, creative, intuitive, 217 00:13:19,480 --> 00:13:23,480 holistic thinking and agents now become some more kind of a 218 00:13:23,480 --> 00:13:26,920 frontal cortex because it organizes, it prioritizes, it 219 00:13:26,920 --> 00:13:29,160 acts. This is extremely powerful, but 220 00:13:29,160 --> 00:13:31,920 of course, you can imagine it introduces quite a big new 221 00:13:31,920 --> 00:13:34,040 complexity. So without what you meant a 222 00:13:34,040 --> 00:13:38,280 strong choice architecture where it's very clear who is leading 223 00:13:38,280 --> 00:13:40,080 when. So when do humans lead? 224 00:13:40,080 --> 00:13:42,440 When is AI leading? And how do they intact? 225 00:13:42,440 --> 00:13:45,960 When you don't follow such an architecture or think about it, 226 00:13:45,960 --> 00:13:49,760 then organizations risk extreme chaos or human paralysis. 227 00:13:49,800 --> 00:13:52,680 So the best companies they defined is upfront. 228 00:13:52,680 --> 00:13:55,400 So they test agents and controlled workflows. 229 00:13:55,400 --> 00:13:58,920 They ensure that decisions are transparent and they also know 230 00:13:58,920 --> 00:14:01,520 that humans need to keep a supervisory role. 231 00:14:01,800 --> 00:14:05,080 And also, yeah, make clear that ethics and and judgement are 232 00:14:05,080 --> 00:14:07,320 essential. So for every case along your, 233 00:14:07,320 --> 00:14:11,040 your life cycle and work workflow or use case, you need 234 00:14:11,040 --> 00:14:14,800 to clearly define when it's fully AI autonomy of guard 235 00:14:14,800 --> 00:14:18,000 rates, when human is in the loop or when it's just humans only. 236 00:14:18,000 --> 00:14:21,720 And if you do this well, it becomes such a multiplier. 237 00:14:21,720 --> 00:14:25,040 You have human creativity, you have great decision quality. 238 00:14:25,040 --> 00:14:28,000 The potential is already visible in our research. 239 00:14:28,000 --> 00:14:31,280 We also question like first, how many companies are experimenting 240 00:14:31,360 --> 00:14:34,680 with agents already? And it's, it's almost 50%, but 241 00:14:34,680 --> 00:14:36,680 just a fraction of those see value. 242 00:14:36,680 --> 00:14:39,320 I think that's also in line with our overall study that few 243 00:14:39,320 --> 00:14:43,520 companies see value, but the share of agentic value from the 244 00:14:43,520 --> 00:14:47,600 total value business value of AI is supposed to double in the 245 00:14:47,600 --> 00:14:50,040 next three years. So it's a, it's a big thing 246 00:14:50,040 --> 00:14:53,520 coming up now you need to act with this responsibly, also 247 00:14:53,520 --> 00:14:56,920 create trust. But then I think the impact will 248 00:14:56,920 --> 00:15:00,000 be will be major and we will all experience this in the next 249 00:15:00,000 --> 00:15:02,760 couple of years. And what I find so intriguing 250 00:15:02,760 --> 00:15:08,840 about including genetic system services or like also like 251 00:15:08,840 --> 00:15:13,840 particular structures into workflows is that it has and 252 00:15:13,840 --> 00:15:17,720 requires A strategic view on the operating model because it's not 253 00:15:17,720 --> 00:15:20,200 a tool. It's like how future built 254 00:15:20,200 --> 00:15:24,520 organizations are defined. And in your expertise, how does 255 00:15:24,520 --> 00:15:28,000 that what really definition of like the AI change and the 256 00:15:28,160 --> 00:15:32,640 interference within organizations links to how 257 00:15:32,640 --> 00:15:38,160 organizations think about trust when it comes to systems, teams 258 00:15:38,160 --> 00:15:44,400 or in the end also the decisions that are taken or influenced and 259 00:15:44,400 --> 00:15:47,280 for employees and like their specific roles. 260 00:15:47,680 --> 00:15:51,560 Yeah, I mean, clearly as AI evolves from currently being 261 00:15:51,560 --> 00:15:55,680 often set a tool to becoming your your business, I think 262 00:15:55,680 --> 00:15:58,320 trust itself needs to be redefined. 263 00:15:58,440 --> 00:16:01,640 I mean, previously and when you have traditional systems or you 264 00:16:01,640 --> 00:16:05,360 had to implement compliance, you had to implement control its But 265 00:16:05,360 --> 00:16:08,480 now in AI driven systems, I think what gives you trust is 266 00:16:08,480 --> 00:16:12,320 also transparency, reliability, but also shared accountability. 267 00:16:12,320 --> 00:16:14,920 True leaders would never ask you, oh, can we trust this 268 00:16:14,920 --> 00:16:16,680 model? They would say, oh, have we 269 00:16:16,680 --> 00:16:21,040 built a system that earns trust from employees, stakeholders, 270 00:16:21,040 --> 00:16:25,480 regulators, everyone, because it controls AI itself and then it 271 00:16:25,480 --> 00:16:28,880 becomes kind of systematic. So I mean decision making 272 00:16:28,880 --> 00:16:31,480 authority. So how you make decision must 273 00:16:31,520 --> 00:16:34,360 align with with risk exposure. I mean, when when there's a 274 00:16:34,360 --> 00:16:38,680 likelihood that you have a massive irreversible catastrophe 275 00:16:38,800 --> 00:16:42,440 of impact from, from a false decision on AI, then you can't 276 00:16:42,440 --> 00:16:45,040 really fully replace human judgement with AI. 277 00:16:45,040 --> 00:16:48,480 Nowadays, whatever you do with AI, the tech stack must be 278 00:16:48,480 --> 00:16:50,480 highly secure against cyber threats. 279 00:16:50,480 --> 00:16:53,600 Otherwise you have yeah, all these problems from infringement 280 00:16:53,840 --> 00:16:57,800 and provency risk, etcetera. And now the more you scale these 281 00:16:57,800 --> 00:17:01,960 systems and the more you build ecosystem, the ETA connectivity 282 00:17:02,000 --> 00:17:04,839 increases exponentially. So everything becomes much more 283 00:17:04,839 --> 00:17:07,079 vulnerable. So you need to even have more 284 00:17:07,079 --> 00:17:10,440 secure operations. So training is important. 285 00:17:10,560 --> 00:17:13,800 And I have also a very interesting example because I 286 00:17:13,800 --> 00:17:16,520 mean, OK, trust, if you think about trust for AI, I mean, we 287 00:17:16,520 --> 00:17:20,960 all use these tools and we get outputs and you initially try or 288 00:17:20,960 --> 00:17:24,200 you initially learn to trust it. To which degree do you challenge 289 00:17:24,200 --> 00:17:27,000 what you what you receive and do your challenges every time or 290 00:17:27,000 --> 00:17:28,960 just sometimes? So so when do you think you need 291 00:17:28,960 --> 00:17:31,680 to challenge an output? And when we did our study at one 292 00:17:31,680 --> 00:17:35,960 point in time, I just put the entire data into our models. 293 00:17:35,960 --> 00:17:38,440 And yeah, of course we have protected environment. 294 00:17:38,440 --> 00:17:42,480 It was not public, but I simply wanted AI to give me hypothesis 295 00:17:42,480 --> 00:17:46,920 to give me great insights. And I got 5 very stunning 296 00:17:47,000 --> 00:17:51,160 out-of-the-box, non obvious ideas from all the data we had. 297 00:17:51,160 --> 00:17:54,080 And I was really impressed. I mean, one of the fives turned 298 00:17:54,080 --> 00:17:56,880 out to be like, like fully, fully off, like wrong 299 00:17:58,160 --> 00:18:00,400 constraints and. But the other four were 300 00:18:00,400 --> 00:18:01,880 brilliant. And then I wanted to say, OK, 301 00:18:01,880 --> 00:18:04,120 let's be mindful, let's challenge it. 302 00:18:04,120 --> 00:18:07,000 Let's ask the algorithm to give me the Python scripts. 303 00:18:07,000 --> 00:18:09,920 Which data was used? Which assumptions did you take? 304 00:18:09,920 --> 00:18:13,560 And it turned out after a while, the two of these four very 305 00:18:13,560 --> 00:18:15,680 impressive insights were completely wrong. 306 00:18:15,960 --> 00:18:18,280 Like, like the algorithm was cycling me. 307 00:18:18,400 --> 00:18:22,360 I use variable epsilon, but they use variable X that there was a 308 00:18:22,360 --> 00:18:24,640 mathematical flaw in a statistical model. 309 00:18:24,880 --> 00:18:27,280 So so that that was very eye opening. 310 00:18:27,280 --> 00:18:31,080 So you still need to do your own homework and create your own 311 00:18:31,080 --> 00:18:33,480 judgement on what you see. I think that's core. 312 00:18:33,480 --> 00:18:36,080 Also to implement it in the operating model to make it 313 00:18:36,080 --> 00:18:41,200 trustworthy at the end. Your situation just shows how we 314 00:18:41,200 --> 00:18:44,800 collaborate with agents or systems or also the 315 00:18:45,040 --> 00:18:46,600 infrastructures that we've built. 316 00:18:46,680 --> 00:18:52,280 And now we don't act like a loan and we want to integrate it into 317 00:18:52,280 --> 00:18:55,800 the ways that we work and want to make sure that the infos that 318 00:18:55,800 --> 00:18:58,880 we share are accurate. I do think that it comes a lot 319 00:18:58,880 --> 00:19:02,320 to the kind of orchestration that we choose between humans, 320 00:19:02,320 --> 00:19:05,320 data and like agents that we incorporate. 321 00:19:05,320 --> 00:19:11,640 And if you now translate it into your like expertise and the 322 00:19:11,640 --> 00:19:14,680 experience that you've made, what does good orchestration 323 00:19:14,680 --> 00:19:17,480 look like? And what happens when 324 00:19:17,560 --> 00:19:21,720 orchestrations fail in terms of connecting humans, data and 325 00:19:21,720 --> 00:19:26,040 agents overall? I mean, I mean AI, as I said, 326 00:19:26,040 --> 00:19:30,200 they genetic systems are super powerful and they but they, as I 327 00:19:30,200 --> 00:19:32,480 described, can never really work in full isolation. 328 00:19:32,520 --> 00:19:34,600 It's the same with people. People also don't work in 329 00:19:34,600 --> 00:19:35,440 isolation. Yeah. 330 00:19:35,440 --> 00:19:39,800 So the re value emerges from the entire orchestration of humans 331 00:19:39,800 --> 00:19:44,120 data working as one team. And I think in in the well run 332 00:19:44,120 --> 00:19:46,880 organizations, it's all about data at the end. 333 00:19:46,880 --> 00:19:50,800 So, so agents use the data to act, to plan and humans set the 334 00:19:50,800 --> 00:19:54,320 contracts at the context and boundaries and, and what they 335 00:19:54,320 --> 00:19:56,400 want to achieve. Yeah, when this orchestration 336 00:19:56,400 --> 00:19:59,600 fails, you experience the opposite like I, I experienced 337 00:19:59,600 --> 00:20:01,520 with my example. You get either, I mean, with 338 00:20:01,560 --> 00:20:04,280 agents, you probably lose accountability or you get 339 00:20:04,280 --> 00:20:07,920 conflicting outputs, you get wrong decisions. 340 00:20:08,000 --> 00:20:10,680 Maybe the work is being duplicated, which is even a big 341 00:20:10,680 --> 00:20:12,400 thing. Like you ask agents to do 342 00:20:12,400 --> 00:20:15,760 something and you never really when they go in cycles and then 343 00:20:15,760 --> 00:20:18,480 suddenly you burn a lot of tokens and it becomes very 344 00:20:18,480 --> 00:20:20,080 costly. So you need to also make sure 345 00:20:20,080 --> 00:20:23,120 that this is under control. And that's why. 346 00:20:23,480 --> 00:20:28,320 The MIT published once a great article about intelligence 347 00:20:28,320 --> 00:20:32,480 choice architectures and here they made pretty clear that you 348 00:20:32,480 --> 00:20:36,240 need to design to provide decision makers with meaningful 349 00:20:36,240 --> 00:20:38,560 choices. So the agents or the algorithms, 350 00:20:38,560 --> 00:20:41,920 they give you choices that they also give you or predict the 351 00:20:41,920 --> 00:20:43,400 outcome of every of these choices. 352 00:20:43,400 --> 00:20:45,960 What does it mean? And then you interact with them. 353 00:20:46,000 --> 00:20:48,160 If you have humans in the loop and then you improve these 354 00:20:48,160 --> 00:20:50,720 choices. And then you also see how these 355 00:20:50,720 --> 00:20:54,720 choices interconnect with maybe other connections you don't even 356 00:20:54,720 --> 00:20:56,520 know about. Because very often you have 357 00:20:56,520 --> 00:20:59,240 these models and you say, hey, let's do XY that, but you don't 358 00:20:59,240 --> 00:21:02,200 know how this interferes with other parts of the systems. 359 00:21:02,280 --> 00:21:04,840 And eventually you have these great choices, which enables 360 00:21:04,840 --> 00:21:07,560 humans to make much better decisions. 361 00:21:07,560 --> 00:21:11,080 So these agents, in short, can never work alone. 362 00:21:11,160 --> 00:21:16,240 So we we all become one system to a certain degree and need to 363 00:21:16,240 --> 00:21:19,560 make sure that designing these architectures becomes kind of a 364 00:21:19,560 --> 00:21:23,200 new management discipline. It's not just the CEO or the CFO 365 00:21:23,200 --> 00:21:26,200 in the past where design is, we really need to design 366 00:21:26,320 --> 00:21:29,360 intelligent governance architectures around agents. 367 00:21:29,360 --> 00:21:34,800 Now and like digging into the report that you've published and 368 00:21:34,920 --> 00:21:41,720 looking at the stats there, it's shown that only are like over 70 369 00:21:41,800 --> 00:21:46,320 percent of companies actually report unmanaged AI security 370 00:21:46,320 --> 00:21:49,880 risks exactly in those systems or potentially also in the way 371 00:21:49,880 --> 00:21:53,120 that they integrate AI into their daily work whatsoever. 372 00:21:53,120 --> 00:21:57,120 And what do you think? Like is this a trust issue or is 373 00:21:57,120 --> 00:22:00,400 it a leadership issue? Is it like potentially also 374 00:22:00,400 --> 00:22:02,400 responsibility issue that we have? 375 00:22:02,400 --> 00:22:05,760 Or is it just a governance failure because people think 376 00:22:06,040 --> 00:22:09,680 governance is boring or like costly complex to include? 377 00:22:09,720 --> 00:22:12,480 It's interesting if if a leadership issue is always a 378 00:22:12,480 --> 00:22:15,400 trust issue or vice versa. I think leadership is also 379 00:22:15,400 --> 00:22:18,480 challenged at times. So sometimes if you lose trust 380 00:22:18,480 --> 00:22:20,640 in something, it's not even a leadership to blame. 381 00:22:20,640 --> 00:22:24,200 It's maybe the complexity that's that's having much faster 382 00:22:24,200 --> 00:22:26,840 psychic times at the moment. I mean, yes, you're right. 383 00:22:26,840 --> 00:22:30,800 More than 70% of companies have reported that their AI security 384 00:22:30,800 --> 00:22:32,960 risks are not managed. It's very interesting that 385 00:22:32,960 --> 00:22:36,920 almost 80% of companies, they use a third party AI tools, of 386 00:22:36,920 --> 00:22:40,520 course, and not everyone buys or builds their LLMS themselves. 387 00:22:40,560 --> 00:22:44,560 They, we all use third parties and many even know how many 388 00:22:44,560 --> 00:22:46,680 tools they are using. That's also a problem with 389 00:22:46,680 --> 00:22:49,640 shadow AI nowadays that, you know, in the past with shadow 390 00:22:49,680 --> 00:22:53,200 IT, people buy their own tools. And now with AI business units 391 00:22:53,200 --> 00:22:56,120 buy their own AI tools and, and automatively. 392 00:22:56,240 --> 00:23:00,200 And there was a research and an experiment where we realized 393 00:23:00,200 --> 00:23:05,080 that almost 50% / 50% of AI failures originate from these 394 00:23:05,080 --> 00:23:07,720 third party tools. That's not surprising, but many 395 00:23:07,720 --> 00:23:10,720 companies still don't, don't mitigate or manage these risks. 396 00:23:10,760 --> 00:23:13,760 So some of your own risk management ends at your own 397 00:23:13,760 --> 00:23:16,200 company, but you're not really considering third parties, 398 00:23:16,200 --> 00:23:19,760 especially I and nowadays agents and multi agent architectures. 399 00:23:19,760 --> 00:23:23,240 So you really open Pandora's box and then people lose confidence 400 00:23:23,240 --> 00:23:26,400 in the system and then suddenly you have leaders who say we need 401 00:23:26,400 --> 00:23:30,600 to be super compliant here and otherwise you, I don't know, get 402 00:23:30,600 --> 00:23:33,400 a big issue. Yeah, you're personally reliable 403 00:23:33,400 --> 00:23:37,160 if something goes wrong and then people don't use AI anymore, and 404 00:23:37,160 --> 00:23:39,680 then what started with a technical issue or wrong risk 405 00:23:39,680 --> 00:23:41,720 management becomes a completely trust issue. 406 00:23:41,960 --> 00:23:46,120 People don't adopt. So it's why the more you scale, 407 00:23:46,480 --> 00:23:49,960 the more you need to make sure that you design. 408 00:23:50,000 --> 00:23:53,200 And that's also core in the book we wrote together a couple of 409 00:23:53,360 --> 00:23:56,400 years or two, 2-3 years, Yeah, two years ago, two years ago. 410 00:23:56,400 --> 00:24:02,320 That exactly time flies that you can't just design a process and 411 00:24:02,320 --> 00:24:05,480 then at the end when it's done, think about, oh, what do we do 412 00:24:05,480 --> 00:24:09,360 with security, privacy, ethics? That needs to be core in the 413 00:24:09,360 --> 00:24:13,720 design when you start designing these things that at the end, if 414 00:24:13,720 --> 00:24:15,880 you run them, this topic does not pop up. 415 00:24:15,880 --> 00:24:19,120 So that's that's very important. So governance is not about 416 00:24:19,120 --> 00:24:22,520 boring bureaucracy. I think it's the foundation when 417 00:24:22,520 --> 00:24:25,960 well done and to create confidence and only then people 418 00:24:25,960 --> 00:24:29,040 use it and you adopt and you innovate and at the end still 419 00:24:29,040 --> 00:24:31,720 stays safe. And one of the key pillars that 420 00:24:31,720 --> 00:24:35,760 you've also worked on was the Future Build Playbook. 421 00:24:35,760 --> 00:24:40,400 And within it, there were like 5 highlights that I found like 422 00:24:40,400 --> 00:24:44,280 particularly worthwhile to dive into because they are also 423 00:24:44,280 --> 00:24:47,760 positioned as 5 strategic imperatives that we need to look 424 00:24:47,760 --> 00:24:50,240 into. The multi year ambition, the 425 00:24:50,240 --> 00:24:55,440 workflow reinvention and AI first model talent and data. 426 00:24:55,440 --> 00:24:59,680 And if you had to pick one that leader is consistently like 427 00:24:59,680 --> 00:25:02,320 literally consistency underestimate, which one would 428 00:25:02,320 --> 00:25:05,520 it be and why? I think to give the answer 429 00:25:05,520 --> 00:25:09,840 first, I think the multi year strategic ambition is super 430 00:25:09,840 --> 00:25:12,000 important. However, everything here is 431 00:25:12,000 --> 00:25:16,320 important because the five, because some of them become an 432 00:25:16,320 --> 00:25:18,360 inhibitor later if you don't do this right. 433 00:25:18,400 --> 00:25:21,440 And as I mentioned also before, the ambition is great, but you 434 00:25:21,440 --> 00:25:23,800 also need to make sure not to forget technology and data 435 00:25:23,800 --> 00:25:26,840 because otherwise later you find barriers to scale. 436 00:25:26,840 --> 00:25:30,200 So everything met us. And if you want to become a 437 00:25:30,200 --> 00:25:33,520 future built company, everything needs to be be in place. 438 00:25:33,520 --> 00:25:36,640 Thinking about a playbook. A playbook is something which 439 00:25:36,640 --> 00:25:39,440 also gives you steps. I mean, if you write a playbook, 440 00:25:39,440 --> 00:25:41,120 it's boring. If you say, hey, please have 441 00:25:41,120 --> 00:25:43,720 this thing and things in place, that's not that's maybe a 442 00:25:43,720 --> 00:25:45,200 checklist, but it's not a playbook. 443 00:25:45,200 --> 00:25:46,920 Yeah, you need to create a journey. 444 00:25:46,960 --> 00:25:52,840 So when we also do these studies and know what are the key levels 445 00:25:52,840 --> 00:25:57,440 to success, of course, we also pave or design A journey also. 446 00:25:57,440 --> 00:26:00,440 And also depending on your starting point, where you are, 447 00:26:00,440 --> 00:26:02,680 what is the priority of disrespective stage. 448 00:26:02,760 --> 00:26:05,520 And definitely in the beginning, you need to have a multi year 449 00:26:05,520 --> 00:26:07,440 strategic commitment from the top. 450 00:26:07,440 --> 00:26:10,080 So it's not like her let's do this as a side project. 451 00:26:10,080 --> 00:26:13,000 It needs to be a North Star because because again, if C 452 00:26:13,000 --> 00:26:16,200 levels just say this is another tool, another layer on top on 453 00:26:16,200 --> 00:26:19,760 what we actually do, then you can gain some efficiencies, but 454 00:26:19,880 --> 00:26:24,000 you will never convert your company to be AI 1st and design 455 00:26:24,000 --> 00:26:26,080 your entire business model around it. 456 00:26:26,160 --> 00:26:31,080 But I mean, at the end, I simply see technology being back in, 457 00:26:31,120 --> 00:26:34,160 in, in the very, very, very, very, very game again. 458 00:26:34,200 --> 00:26:37,480 I mean, many years, you know, many years ago, like it was 459 00:26:37,480 --> 00:26:40,120 always IT projects and you had to ask the business. 460 00:26:40,120 --> 00:26:42,680 Then it turned around to be like, hey, this is a business 461 00:26:42,680 --> 00:26:44,720 transformation, but please involve it. 462 00:26:44,760 --> 00:26:47,600 And the best companies, they they combine this. 463 00:26:47,600 --> 00:26:50,600 Yeah, they really have make this a business and tech 464 00:26:50,600 --> 00:26:51,880 transformation. Likewise. 465 00:26:51,880 --> 00:26:55,320 And at the moment, just to give one example, I mean, everyone is 466 00:26:55,320 --> 00:26:57,760 now having these big SAP transformations. 467 00:26:57,760 --> 00:27:01,240 Yeah, 2S for Hannah. And this is not a tech 468 00:27:01,240 --> 00:27:04,440 technology update here. This is so significant because 469 00:27:04,440 --> 00:27:08,120 you turn around every business unit, every process and 470 00:27:08,120 --> 00:27:10,680 everything is challenged by its value. 471 00:27:10,680 --> 00:27:13,560 And it's costly. You need to prove the value of 472 00:27:13,560 --> 00:27:15,400 it. And at the end, it's also all 473 00:27:15,400 --> 00:27:19,200 about redesigning your data. Tech is also a business 474 00:27:19,200 --> 00:27:21,680 transformation, not only business transformation. 475 00:27:21,680 --> 00:27:23,720 So take it like it's it goes hand in hand. 476 00:27:23,840 --> 00:27:28,200 And the hand in hand, side by side collaboration is something 477 00:27:28,200 --> 00:27:30,520 that you've also in like your work. 478 00:27:30,960 --> 00:27:34,560 So the study said you've published emphasized quite a bit 479 00:27:34,600 --> 00:27:39,480 because sharing this exact ownership and also combining it 480 00:27:39,480 --> 00:27:44,120 to mature like as organizations and with the teams is quite hard 481 00:27:44,120 --> 00:27:47,680 actually. And I also perceive this as one 482 00:27:47,680 --> 00:27:51,920 of the hindrances because mutual understanding is super hard to 483 00:27:51,920 --> 00:27:54,800 establish. And could you trade a bit more 484 00:27:54,800 --> 00:27:59,680 on why that is and also how organizations can practically 485 00:27:59,680 --> 00:28:03,280 build that shared ownership into their governance set up? 486 00:28:03,280 --> 00:28:06,720 Yeah, yeah, pretty clear. I mean, shared ownership sounds 487 00:28:06,720 --> 00:28:10,840 initially logical, but I mean we know the word of reporting 488 00:28:10,840 --> 00:28:16,240 lines, the word of budgets, the word of politics, and also the 489 00:28:16,240 --> 00:28:18,600 word of, I must say, key accountabilities. 490 00:28:18,680 --> 00:28:21,160 I mean the tension. If we think a couple of years 491 00:28:21,160 --> 00:28:24,920 back, business teams always wanted to have the IT solution 492 00:28:24,920 --> 00:28:26,640 that best fits the individual problem. 493 00:28:26,640 --> 00:28:30,080 So they give demands to ITIT builds it and eventually you 494 00:28:30,080 --> 00:28:32,720 have a super heterogeneic architecture because you 495 00:28:32,720 --> 00:28:34,520 basically serve 100 different demands. 496 00:28:34,520 --> 00:28:39,200 So IT at one point was focused to ensure stability and security 497 00:28:39,200 --> 00:28:42,920 and also scalability. And unfortunately, IT was also 498 00:28:42,920 --> 00:28:46,600 pretty much squeezed like a like a lemon over years to save more 499 00:28:46,600 --> 00:28:48,840 and more costs. Yeah, so some of them are really 500 00:28:48,880 --> 00:28:53,080 burdened by years of of of of, of, of cost savings. 501 00:28:53,080 --> 00:28:56,480 Yeah. So then IT suddenly got the huge 502 00:28:56,480 --> 00:28:58,280 mandate. OK, we need to have a digital 503 00:28:58,280 --> 00:28:59,800 transformation. Can you do this on top? 504 00:28:59,800 --> 00:29:01,760 Wonderful. So on the one hand, with your 505 00:29:01,760 --> 00:29:05,480 tiny budget, you need to be the innovator of your organization 506 00:29:05,480 --> 00:29:07,760 and innovate rapidly. On the other hand, you need to 507 00:29:07,760 --> 00:29:11,280 make sure that your architecture is not becoming spaghetti or how 508 00:29:11,280 --> 00:29:13,240 we called it sometimes. So it wasn't easy. 509 00:29:13,320 --> 00:29:16,360 And then you suddenly created these new roads like there's a 510 00:29:16,360 --> 00:29:19,840 chief digital officer now and there was often tangents because 511 00:29:19,840 --> 00:29:22,960 like both had their different targets, but it was not without 512 00:29:22,960 --> 00:29:25,240 any conflict. Then it was clear everything has 513 00:29:25,240 --> 00:29:28,280 to be a business transformation. We need to put a business person 514 00:29:28,280 --> 00:29:30,520 in charge of everything. And this was great exciting 515 00:29:30,880 --> 00:29:34,600 because suddenly you can you can also have great communication 516 00:29:34,600 --> 00:29:37,840 how AI and digital turns around your entire business. 517 00:29:37,840 --> 00:29:40,840 And then you try to experiment a lot, but suddenly you have 518 00:29:40,880 --> 00:29:43,920 hundreds of great pilots. So we always call this thousand 519 00:29:43,920 --> 00:29:46,320 flowers blooming. So we basically plant an entire 520 00:29:46,320 --> 00:29:49,000 garden, but all the roses die at the end. 521 00:29:49,040 --> 00:29:51,400 So and then there's the moment of truth where you need to 522 00:29:51,400 --> 00:29:54,720 think, why is this the case? And then you realize that maybe 523 00:29:54,720 --> 00:29:59,040 tech and digital and data must be part of this entire 524 00:29:59,040 --> 00:30:01,400 transformation. Then it brings to clearly back 525 00:30:01,400 --> 00:30:03,840 to what I'm saying in the beginning is that you need to be 526 00:30:03,840 --> 00:30:05,440 called a sometimes two in the box. 527 00:30:05,480 --> 00:30:10,200 So all these topics need to have a business and an IT digital 528 00:30:10,280 --> 00:30:13,400 data counterpart. And if we, if I may share 1 529 00:30:13,720 --> 00:30:17,520 figure from our report, we ask companies how they organize that 530 00:30:17,520 --> 00:30:19,120 and the companies which are future bills. 531 00:30:19,120 --> 00:30:22,600 So the top 5%, they have 50% more. 532 00:30:22,600 --> 00:30:25,680 This organization is set up of having these two in the box 533 00:30:25,680 --> 00:30:27,640 ownerships. And I don't mean like just 534 00:30:27,640 --> 00:30:31,280 collaborating, I mean shared responsibility, shared targets, 535 00:30:31,320 --> 00:30:33,320 shared accountability, shared budgets. 536 00:30:33,320 --> 00:30:36,000 So, so really not not just governance on paper, but 537 00:30:36,000 --> 00:30:38,480 governance established, which is a huge different. 538 00:30:38,480 --> 00:30:41,480 Actually, I've I've over years of consulting, you can't imagine 539 00:30:41,480 --> 00:30:43,400 how much governance I saw on paper. 540 00:30:43,400 --> 00:30:45,280 But the reality was completely different. 541 00:30:45,280 --> 00:30:48,360 So this time it's for real. Yeah, you need to anchor it to 542 00:30:48,360 --> 00:30:50,560 make it happen. I would love to dive into that, 543 00:30:50,560 --> 00:30:54,360 but I do think that then the time would fly even more than it 544 00:30:54,360 --> 00:30:57,880 already does so. But for now, I do think one of 545 00:30:57,880 --> 00:31:00,520 the aspects that I want to piggyback on is the 546 00:31:00,600 --> 00:31:03,560 consciousness of the leaders that they sometimes have when 547 00:31:03,560 --> 00:31:07,480 either it comes to the shared ownership or the established and 548 00:31:07,480 --> 00:31:10,120 framework that is required to bring that to life. 549 00:31:10,120 --> 00:31:14,640 Because there is this huge word that everyone kind of wants to 550 00:31:14,640 --> 00:31:19,480 avoid because it's regulation. And we all say that it's hard 551 00:31:19,480 --> 00:31:22,680 for regulation to catch up before it's scaling AI. 552 00:31:22,760 --> 00:31:26,800 And from your perspectives, what actually are the costs of 553 00:31:26,800 --> 00:31:29,440 waiting? And what kind of leadership 554 00:31:29,440 --> 00:31:32,920 mindset do we need instead when we do want to progress? 555 00:31:33,160 --> 00:31:36,080 I mean, on the one hand, I have this very rational mind. 556 00:31:36,080 --> 00:31:40,520 I like innovation. I like rapid innovation, speed, 557 00:31:40,520 --> 00:31:43,400 getting results fast. And on the other hand, I also 558 00:31:43,400 --> 00:31:46,720 have a half a young daughter and I want to grow up in a in a very 559 00:31:46,840 --> 00:31:50,240 safe and protected environment and avoid the robotic 560 00:31:50,240 --> 00:31:51,160 apocalypse. Yeah. 561 00:31:51,160 --> 00:31:54,400 So I think it's it's it's it's two ways how to look at this 562 00:31:54,400 --> 00:31:56,080 topic. First, the rational point. 563 00:31:56,120 --> 00:31:58,280 I mean, if you wait, then you can wait forever. 564 00:31:58,280 --> 00:32:01,800 You slow down innovation and your competitor US may progress. 565 00:32:01,800 --> 00:32:04,000 And you know, we also have global competition. 566 00:32:04,000 --> 00:32:06,320 It's not just your competitor around the corner, it's a 567 00:32:06,320 --> 00:32:07,920 competitor at the end of the world. 568 00:32:07,960 --> 00:32:11,640 So you need to you need to also progress and also own your turf 569 00:32:11,640 --> 00:32:13,680 and keep your competitive advantage. 570 00:32:13,680 --> 00:32:16,240 The longer you wait, you also create more. 571 00:32:16,240 --> 00:32:18,440 We call this tech debt. I mean, the big challenge for 572 00:32:18,440 --> 00:32:21,560 companies nowadays is that they are still adopting these new 573 00:32:21,560 --> 00:32:23,760 technologies, but they still keep their current tech 574 00:32:23,760 --> 00:32:26,440 architectures. And this is very, very, very 575 00:32:26,440 --> 00:32:28,160 costly. And when you wait too long, you 576 00:32:28,160 --> 00:32:30,440 also don't benefit from learning cycle. 577 00:32:30,600 --> 00:32:34,200 So at the end, those companies which, which shape guard rates 578 00:32:34,200 --> 00:32:37,160 now, which tests responsibly, which also take personal 579 00:32:37,160 --> 00:32:40,640 ownership for the topics and bid readiness, they, they, they jump 580 00:32:40,640 --> 00:32:43,240 ahead on the curve. So I think you shouldn't wait 581 00:32:43,240 --> 00:32:45,320 too long. You should bid the capabilities, 582 00:32:45,320 --> 00:32:48,120 but do it in a responsible bid manner from the heart. 583 00:32:48,120 --> 00:32:51,600 And the other perspective is, I mean, I hope and that's leaving 584 00:32:51,600 --> 00:32:53,640 now. Maybe, maybe the, the current 585 00:32:53,640 --> 00:32:55,960 discussion a bit what is best for humanity? 586 00:32:55,960 --> 00:32:59,560 I mean, we are all humans now. We are not trans humans yet who 587 00:32:59,640 --> 00:33:02,720 send their DNA to space. So we are still all together 588 00:33:02,720 --> 00:33:05,000 here on this planet. And I think we should share a 589 00:33:05,000 --> 00:33:10,360 common goal and be all aligned that AI should everyone serve as 590 00:33:10,360 --> 00:33:12,080 a greater good. I mean, the tension is is 591 00:33:12,080 --> 00:33:14,200 immense. I mean, think about tackling 592 00:33:14,200 --> 00:33:17,680 climate change, think about curing diseases, education is 593 00:33:17,680 --> 00:33:20,160 available, quality of life gets better. 594 00:33:20,160 --> 00:33:22,200 So I recently thought about one word. 595 00:33:22,200 --> 00:33:25,480 And I mean, no one wants to be remembered as the Oppenheimer of 596 00:33:25,480 --> 00:33:28,400 AI who created then the beast that was unleashed. 597 00:33:28,400 --> 00:33:30,520 And years later, you have the big regret. 598 00:33:30,640 --> 00:33:33,760 So I think both, both perspectives fit together. 599 00:33:33,760 --> 00:33:37,040 I'm a strong advocate for fast innovation, learning progress, 600 00:33:37,040 --> 00:33:39,320 but it's important to define red zones. 601 00:33:39,320 --> 00:33:42,040 So something which like, I mean, we, we were pretty clear with 602 00:33:42,040 --> 00:33:46,000 GDPR for data and I think wasn't easy, but I think we also get 603 00:33:46,000 --> 00:33:47,840 something good out of it. And I think we should do the 604 00:33:47,840 --> 00:33:50,680 same with AI. And one thing which is important 605 00:33:50,680 --> 00:33:53,320 to share, maybe as a last insight from the study 606 00:33:53,320 --> 00:33:57,160 responsibly, AI is one topic we survey and assess in our 607 00:33:57,160 --> 00:33:59,880 studies. And interestingly, in the 608 00:33:59,880 --> 00:34:03,120 beginning, the topic didn't really pop up that much like you 609 00:34:03,120 --> 00:34:06,360 should need to focus on other topics first to digitize fast 610 00:34:06,360 --> 00:34:09,440 and bid AI. But those companies and we do 611 00:34:09,440 --> 00:34:13,719 this crazy multivariate analysis with cause and models etcetera. 612 00:34:13,719 --> 00:34:17,320 I spare you the details, but we have seen that from the third 613 00:34:17,320 --> 00:34:21,040 stage to the fourth stage to become a future bid responsibly 614 00:34:21,040 --> 00:34:26,040 AI was the most significant inhibitor to become future bid 615 00:34:26,040 --> 00:34:29,239 if you haven't really designed it until the stage. 616 00:34:29,239 --> 00:34:32,600 So you can neglect it for, but it will backfire at the end when 617 00:34:32,600 --> 00:34:35,400 you really want to scale. And it was quite comforting for 618 00:34:35,400 --> 00:34:37,920 me. So there's clear proven value in 619 00:34:37,920 --> 00:34:40,199 having responsibility AI. Amazing. 620 00:34:40,239 --> 00:34:44,000 I think we've also touched upon that in our like book that we've 621 00:34:44,000 --> 00:34:47,600 published 10 moral questions and there's like the whole 622 00:34:47,600 --> 00:34:50,080 background on the economical part as well. 623 00:34:50,080 --> 00:34:53,040 So I'll link the link to the background information on the 624 00:34:53,040 --> 00:34:55,679 study in the book. Also in the show notes. 625 00:34:55,679 --> 00:34:59,960 Unfortunately, we already come to close with like today's 626 00:34:59,960 --> 00:35:02,480 episode and also with our conversation. 627 00:35:02,480 --> 00:35:06,080 But before we just say goodbye, I do want to share some moments 628 00:35:06,080 --> 00:35:08,440 with you. They're so-called in between 629 00:35:08,440 --> 00:35:10,920 moments. I use them with every guest as 630 00:35:10,920 --> 00:35:14,400 quick, intuitive reflections. And I would love for you to 631 00:35:14,400 --> 00:35:17,280 answer in like a word or a sentence, but pretty short. 632 00:35:17,280 --> 00:35:22,440 And I want to start on what are how trust in AI feels like for 633 00:35:22,440 --> 00:35:24,600 you. Sometimes I wish I could be AI 634 00:35:24,600 --> 00:35:28,360 for a day. Yeah, it's a bit. 635 00:35:28,800 --> 00:35:33,280 You know, it would be great if, if, if AI could feel like humans 636 00:35:33,280 --> 00:35:35,640 feel and humans could feel like AI feel. 637 00:35:35,640 --> 00:35:38,240 So I don't know if this is achievable, but that would give 638 00:35:38,240 --> 00:35:41,040 me trust. And what does Europe need more 639 00:35:41,040 --> 00:35:43,600 of? I think we we can have a bit 640 00:35:43,600 --> 00:35:47,400 more self esteem and trust our own capabilities, but what we 641 00:35:47,400 --> 00:35:51,800 really need is pragmatism. And what does AI leadership 642 00:35:51,800 --> 00:35:54,400 require? Ah, you touched courage before. 643 00:35:54,400 --> 00:35:57,800 I think courage is important, but courage defined with 644 00:35:57,800 --> 00:36:01,840 integrity, not not just being radical, but being courage, 645 00:36:02,000 --> 00:36:06,200 courageous, but in a mindful matter and having also some 646 00:36:06,200 --> 00:36:08,920 patience I think. And what is the most overlooked 647 00:36:08,920 --> 00:36:13,160 skill in transformation? I think people driven 648 00:36:13,240 --> 00:36:17,440 innovation, yeah, I think enable people, make them curious, give 649 00:36:17,440 --> 00:36:21,200 them cool tools and let them do really use people to innovate. 650 00:36:21,400 --> 00:36:25,360 Enough that last one. What does future built mean? 651 00:36:25,480 --> 00:36:30,360 Those future built companies, I have seen a very laser focus or 652 00:36:30,360 --> 00:36:33,560 not truly matters, but they are also very bold and execution. 653 00:36:33,560 --> 00:36:36,120 I sometimes envy them. I also sometimes wish I would be 654 00:36:36,120 --> 00:36:38,440 more focused and what probably I am. 655 00:36:38,440 --> 00:36:40,200 But you know, it's sometimes stunning. 656 00:36:40,240 --> 00:36:43,600 It it's, I mean, think about Steve Jobs in the past. 657 00:36:43,600 --> 00:36:48,640 He had always 100% clear knowledge what he wants to do 658 00:36:48,640 --> 00:36:51,120 and what matters. That's that's that Champions 659 00:36:51,120 --> 00:36:51,600 League. Yeah. 660 00:36:52,560 --> 00:36:55,920 Indeed it is. And I want to close with one 661 00:36:55,920 --> 00:37:01,640 question that I would love your like view on point of view or 662 00:37:01,640 --> 00:37:06,480 maybe also outlook, because if trust is a missing currency 663 00:37:06,480 --> 00:37:10,680 between the ambition that we have and the execution on like 664 00:37:10,680 --> 00:37:14,800 the clarity and sophistication on implementation, what does it 665 00:37:14,800 --> 00:37:20,120 mean in 2025 to be a trust builder in AI transformation? 666 00:37:20,160 --> 00:37:21,200 I. Mean we wrote this book 667 00:37:21,200 --> 00:37:24,440 together, yeah, about 10 more questions and to test the titles 668 00:37:24,440 --> 00:37:28,440 like how to design take and AI responsibly. 669 00:37:28,440 --> 00:37:31,600 And there are a couple of aspects which are really 670 00:37:31,600 --> 00:37:34,440 important to establish trust in AI. 671 00:37:34,560 --> 00:37:37,400 And I mean, as we said, I mean make sure it's beneficial to 672 00:37:37,400 --> 00:37:39,680 everyone. Like it has to be a common 673 00:37:39,680 --> 00:37:42,560 better good for your customers, stakeholders, your workforce. 674 00:37:42,560 --> 00:37:46,040 I think that's important. And as we discussed, make sure 675 00:37:46,040 --> 00:37:50,640 that it's it's humans and AI jointly being in this governance 676 00:37:50,640 --> 00:37:54,080 models and strive to an intended purpose like give, give agents 677 00:37:54,080 --> 00:37:56,080 freedom, but also make sure where freedoms end. 678 00:37:56,120 --> 00:37:59,720 And the last pieces may be coming back to my my being in 679 00:37:59,720 --> 00:38:03,400 favor of people and humans. Educate people, upscale people 680 00:38:03,400 --> 00:38:07,440 and power them and use them to reshape their own workforce and 681 00:38:07,440 --> 00:38:09,960 processes. So then adoption also gets 682 00:38:09,960 --> 00:38:11,320 higher, etcetera. Yeah. 683 00:38:11,320 --> 00:38:14,920 And one thing as we as we finish this interview, I had to laugh 684 00:38:14,920 --> 00:38:19,040 because there's one favorite quote, which I also jumps and 685 00:38:19,040 --> 00:38:22,040 making fun to my wife that can put on my tombstone. 686 00:38:22,040 --> 00:38:26,000 Yeah, I hope there's still many years to go, but I like always 687 00:38:26,000 --> 00:38:28,280 the word do the best and trust the process. 688 00:38:28,320 --> 00:38:31,720 I think we should simply continue, do our best, act 689 00:38:31,720 --> 00:38:35,480 responsibly, innovate and maybe be good people. 690 00:38:36,720 --> 00:38:39,760 That's a big ending. It is. 691 00:38:41,200 --> 00:38:44,840 It's quite a statement to to leave this conversation, but I 692 00:38:44,840 --> 00:38:47,400 love it. Thanks so much for spending time 693 00:38:47,400 --> 00:38:50,360 together and sharing also the background on the report. 694 00:38:50,360 --> 00:38:54,600 I appreciate your time and like the strong finish and look 695 00:38:54,600 --> 00:38:56,400 forward to seeing where that brings us. 696 00:38:56,400 --> 00:38:59,680 But for now, thanks so much to man and we'll speak soon. 697 00:38:59,760 --> 00:39:00,760 Yeah. Thank you. 698 00:39:00,760 --> 00:39:02,200 If it was great. Thanks.