WEBVTT 1 00:00:00.240 --> 00:00:03.160 - We traditionally very much thought in pilots, 2 00:00:03.160 --> 00:00:05.520 but at some point we realized, if you really want 3 00:00:05.520 --> 00:00:07.240 to change things, you need 4 00:00:07.240 --> 00:00:09.680 to take an approach that's scalable. 5 00:00:10.640 --> 00:00:13.600 - Heidelberg Materials is a giant in building materials 6 00:00:13.600 --> 00:00:17.640 and solutions with market-leading positions in cement, 7 00:00:17.640 --> 00:00:19.760 aggregates, and ready-mix concrete. 8 00:00:20.280 --> 00:00:22.360 Dominik von Achten is the CEO 9 00:00:22.360 --> 00:00:24.320 and the driving force 10 00:00:24.320 --> 00:00:27.960 behind future-proofing Heidelberg Materials' business 11 00:00:27.960 --> 00:00:31.040 and accelerating the company's profitable growth path. 12 00:00:31.040 --> 00:00:33.440 Join me as I sit down with Dominik to learn 13 00:00:33.440 --> 00:00:36.480 what it really takes to reinvent the wheel 14 00:00:36.480 --> 00:00:42.120 by pioneering AI at a company that is 150 years old. 15 00:00:52.840 --> 00:00:54.640 Dominik, technology 16 00:00:54.640 --> 00:00:58.280 and sustainability are so central to the impressive results 17 00:00:58.280 --> 00:01:00.320 of Heidelberg Materials, 18 00:01:00.320 --> 00:01:03.360 and you have a very particular approach to growth. 19 00:01:03.360 --> 00:01:04.640 Tell us more about that. 20 00:01:04.640 --> 00:01:07.440 - I think for Heidelberg Materials, 21 00:01:07.440 --> 00:01:09.400 it is first profits, obviously, 22 00:01:09.400 --> 00:01:11.960 and then it's transformation along the two lines 23 00:01:11.960 --> 00:01:15.760 that you were describing, sustainability, AI and automation, 24 00:01:15.760 --> 00:01:16.880 and also technology. 25 00:01:17.800 --> 00:01:21.960 I think we have to secure the first to get to the second. 26 00:01:21.960 --> 00:01:25.200 In both cases, the industry had some challenges in the past, 27 00:01:26.240 --> 00:01:30.600 both in terms of being a big carbon emitter, second in terms 28 00:01:30.600 --> 00:01:33.920 of being the most offline business on the planet. 29 00:01:33.920 --> 00:01:36.680 And where there are challenges, there are big opportunities, 30 00:01:36.680 --> 00:01:38.560 and in Heidelberg, we grab them. 31 00:01:39.400 --> 00:01:42.560 - That's great to hear. Now, you say you were one 32 00:01:42.560 --> 00:01:46.560 of the most offline businesses out there, yet you decided 33 00:01:46.560 --> 00:01:49.720 to take on AI and you decided to do that 34 00:01:49.720 --> 00:01:51.520 with an enterprise-wide approach. 35 00:01:52.520 --> 00:01:55.240 What drove that decision and how is that working now? 36 00:01:55.240 --> 00:01:58.560 - We traditionally very much thought in pilots, 37 00:01:58.560 --> 00:02:00.960 but at some point we realized if you really want 38 00:02:00.960 --> 00:02:02.320 to change things, you need 39 00:02:02.320 --> 00:02:04.920 to take an approach that's scalable. 40 00:02:04.920 --> 00:02:08.760 And for us, we have big pockets along the whole value chain 41 00:02:08.760 --> 00:02:13.320 that are very, very good to be addressed 42 00:02:13.320 --> 00:02:16.880 with automation and especially also AI. 43 00:02:16.880 --> 00:02:21.120 You know, you can start in the quarries, you can then go 44 00:02:21.120 --> 00:02:25.480 to the labs, you can then go to the production sites, 45 00:02:25.480 --> 00:02:30.400 you can then go to logistics, you can then go to 46 00:02:30.400 --> 00:02:32.520 sales and distribution, 47 00:02:32.520 --> 00:02:34.440 and you can then go to the customer 48 00:02:34.440 --> 00:02:36.120 interface, including marketing. 49 00:02:36.120 --> 00:02:39.760 So the good news is there was so much offline 50 00:02:39.760 --> 00:02:40.880 and there is, there are 51 00:02:40.880 --> 00:02:44.600 so many good spots in our value chain that you can address 52 00:02:44.600 --> 00:02:48.640 with AI that really then also gives us the ability 53 00:02:48.640 --> 00:02:51.400 to build a platform and scale it up. 54 00:02:51.400 --> 00:02:54.360 - And if you go further down your value chain 55 00:02:54.360 --> 00:02:58.040 beyond the queries, other examples that you found exciting? 56 00:02:58.040 --> 00:03:00.960 - Yeah. If you then move, for example, into our plants, 57 00:03:00.960 --> 00:03:04.600 we have now in North America created what I call the, 58 00:03:04.600 --> 00:03:07.960 or what we call the Heidelberg Remote Operation Center 59 00:03:07.960 --> 00:03:11.560 where we now out of one location in Dallas run 12 plants 60 00:03:11.560 --> 00:03:13.520 basically across North America. 61 00:03:14.760 --> 00:03:17.680 You have seen this in other industries, it works. 62 00:03:17.680 --> 00:03:20.120 We are now piloting this in our industry. 63 00:03:20.120 --> 00:03:22.120 I think we are the first company piloting 64 00:03:22.120 --> 00:03:24.400 that at scale in North America 65 00:03:24.400 --> 00:03:26.280 and really making sure 66 00:03:26.280 --> 00:03:28.920 that we can actually then also apply this to the rest 67 00:03:28.920 --> 00:03:30.240 of our operations globally. 68 00:03:30.240 --> 00:03:35.160 Again, pilot, but then immediately with a thought to scale. 69 00:03:35.160 --> 00:03:38.120 - Now, Heidelberg Materials is 150 years old. 70 00:03:38.120 --> 00:03:41.320 It's an iconic organization that also has a strong culture. 71 00:03:42.200 --> 00:03:45.320 How do you bring your workforce along? 72 00:03:45.320 --> 00:03:47.360 How do you get them excited? 73 00:03:47.360 --> 00:03:50.680 How do you get them motivated to contribute and embrace AI? 74 00:03:50.680 --> 00:03:54.480 - In Germany and in Europe, we have a big challenge 75 00:03:54.480 --> 00:03:59.480 because as much as the 152 years are great, a great history, 76 00:03:59.520 --> 00:04:03.720 they can become a burden when you don't get the power 77 00:04:03.720 --> 00:04:07.800 and energy to change if your surroundings do change. 78 00:04:07.800 --> 00:04:09.000 And I think there is a lot 79 00:04:09.000 --> 00:04:11.440 of challenges right now out there. 80 00:04:11.440 --> 00:04:14.640 And our task as a leadership team is to really make sure 81 00:04:14.640 --> 00:04:16.440 that we energize 82 00:04:16.440 --> 00:04:20.680 and rally our workforce around this change. 83 00:04:20.680 --> 00:04:24.800 - Now, one of the big convictions we have at BCG is 84 00:04:24.800 --> 00:04:28.640 for any company to be successful in an AI transformation, 85 00:04:28.640 --> 00:04:32.360 and for sure with an enterprise-wide AI transformation, 86 00:04:32.360 --> 00:04:35.920 the CEO needs to set the tone and needs to lead by example. 87 00:04:35.920 --> 00:04:40.640 What are you doing to signal that AI is important 88 00:04:40.640 --> 00:04:42.520 and that you are part of it? 89 00:04:42.520 --> 00:04:45.080 - Well, the good news is, I don't need to signal much. 90 00:04:45.080 --> 00:04:47.640 I'm really convinced and I'm really excited myself. 91 00:04:47.640 --> 00:04:51.920 That's why we have set up a couple of formats also, 92 00:04:51.920 --> 00:04:54.480 what we call, for example, the "CEO AI hour," 93 00:04:54.480 --> 00:04:57.480 where I personally then interact with a couple 94 00:04:57.480 --> 00:04:59.160 of our colleagues around the world 95 00:04:59.160 --> 00:05:02.040 that have done something, built something with AI. 96 00:05:02.040 --> 00:05:04.240 So it's not about an idea, it's about something 97 00:05:04.240 --> 00:05:07.840 that they have done and that I'm trying to scale it up. 98 00:05:07.840 --> 00:05:11.160 That's a little bit... It's a direct interaction, 99 00:05:11.160 --> 00:05:14.480 if I'm honest, across many hierarchy levels 100 00:05:14.480 --> 00:05:17.920 and that is not only comfortable for the organization, 101 00:05:17.920 --> 00:05:19.680 but I think it should show my leadership, 102 00:05:19.680 --> 00:05:22.800 but at the same time should show us that in a company 103 00:05:22.800 --> 00:05:27.800 of 50,000 people, AI has the ability to empower everybody. 104 00:05:28.400 --> 00:05:32.160 And I want to see who is empowered first in the organization 105 00:05:32.160 --> 00:05:34.640 and then try to scale it up. 106 00:05:34.640 --> 00:05:38.040 - 2025 was a record year for Heidelberg Materials. 107 00:05:38.040 --> 00:05:39.800 You lead the industry 108 00:05:39.800 --> 00:05:42.720 and you have bold ambitions going forward. 109 00:05:42.720 --> 00:05:46.600 How does AI play into the whole kind of 110 00:05:46.600 --> 00:05:48.080 financial philosophy 111 00:05:48.080 --> 00:05:50.440 and the plan philosophy for your company? 112 00:05:51.360 --> 00:05:53.520 - Yeah, I think it's clear that we are all trying 113 00:05:53.520 --> 00:05:57.000 to get our arms around what's the exact impact of AI. 114 00:05:57.000 --> 00:05:59.920 I think there is, 115 00:05:59.920 --> 00:06:02.240 the first idea was, there is an efficiency gain 116 00:06:02.240 --> 00:06:03.480 that you can get. 117 00:06:03.480 --> 00:06:05.480 I think there is a matter of speed 118 00:06:05.480 --> 00:06:07.600 and position that comes with AI 119 00:06:07.600 --> 00:06:09.440 that we are probably underestimating. 120 00:06:10.280 --> 00:06:13.400 I think there is an opportunity on the customer side 121 00:06:13.400 --> 00:06:16.120 for much more top-line development that 122 00:06:17.000 --> 00:06:19.200 I think we are also trying to get to. 123 00:06:19.200 --> 00:06:20.960 That's why I'm in the shadow 124 00:06:20.960 --> 00:06:23.840 of AI also pushing the commercial side of things quite a bit 125 00:06:23.840 --> 00:06:26.200 because I strongly believe you can also 126 00:06:27.480 --> 00:06:29.200 differentiate yourself with the help 127 00:06:29.200 --> 00:06:31.560 of AI on the customer front end 128 00:06:31.560 --> 00:06:33.720 and your product development, for example. 129 00:06:33.720 --> 00:06:36.280 So for us, there's a top-line growth 130 00:06:36.280 --> 00:06:38.200 opportunity that sits there. 131 00:06:38.200 --> 00:06:40.040 And I think then we need to see a little bit 132 00:06:40.040 --> 00:06:43.080 how the cost side develops, whether you can really, 133 00:06:44.000 --> 00:06:47.480 it's the idea that it will reduce a lot 134 00:06:47.480 --> 00:06:49.080 of jobs, I'm not so sure. 135 00:06:49.080 --> 00:06:52.200 I think probably it will give the existing jobs more 136 00:06:52.200 --> 00:06:56.360 abilities to work side by side with an AI agent to 137 00:06:56.360 --> 00:06:59.560 basically leverage our existing workforce. 138 00:06:59.560 --> 00:07:01.120 And then I think the big question 139 00:07:01.120 --> 00:07:04.080 that is still out there is, what's the cost 140 00:07:04.080 --> 00:07:07.720 that AI will also add to the P&L at the end? 141 00:07:07.720 --> 00:07:11.200 Do we get a bigger efficiency or speed 142 00:07:11.200 --> 00:07:15.040 or something increase versus the additional cost we add? 143 00:07:15.040 --> 00:07:18.400 - Dominik, Heidelberg Materials embraces different 144 00:07:18.400 --> 00:07:20.200 flavors of AI. 145 00:07:20.200 --> 00:07:22.480 Obviously predictive AI, which you have been doing 146 00:07:22.480 --> 00:07:24.000 for a long time already. 147 00:07:25.240 --> 00:07:29.440 Generative AI, agentic AI. Now comes physical AI. 148 00:07:30.480 --> 00:07:32.640 How do you think about those different arenas 149 00:07:32.640 --> 00:07:35.720 and where do things go from here? 150 00:07:35.720 --> 00:07:38.840 - I think, for example, in our ready-mix plants, 151 00:07:38.840 --> 00:07:40.600 you typically have mixed designs. 152 00:07:40.600 --> 00:07:44.360 So you have to imagine this like a recipe, 153 00:07:45.320 --> 00:07:47.600 and for cooking almost, 154 00:07:47.600 --> 00:07:51.000 there are hundreds if not thousands per plant, recipes. 155 00:07:51.000 --> 00:07:54.400 What we have done is to connect this physical world 156 00:07:54.400 --> 00:07:58.280 and use the lever of AI is to 157 00:07:58.280 --> 00:08:02.960 collect 180,000 historical recipes that we have mixed 158 00:08:02.960 --> 00:08:07.200 around the world and are using these 180,000 recipes 159 00:08:07.200 --> 00:08:10.400 to now train our models in order to come up 160 00:08:10.400 --> 00:08:14.120 with an industry-wide and company-wide platform. 161 00:08:14.120 --> 00:08:17.000 That's why we have made also some investments in the US into 162 00:08:17.000 --> 00:08:18.440 software businesses. 163 00:08:19.560 --> 00:08:23.640 Giatec Scientific, by the way, they are using these amount 164 00:08:23.640 --> 00:08:25.840 of recipes to train their algorithms 165 00:08:25.840 --> 00:08:30.840 to then improve quality assessment, safety behavior, 166 00:08:30.880 --> 00:08:35.680 content development of those mixes, 167 00:08:35.680 --> 00:08:38.160 marginal optimization of those mixes. 168 00:08:38.160 --> 00:08:39.200 So there are a lot of levers 169 00:08:39.200 --> 00:08:40.560 that you can play if you connect 170 00:08:40.560 --> 00:08:42.840 what the physical world has given you 171 00:08:42.840 --> 00:08:45.920 and what what AI can do in order to leverage 172 00:08:45.920 --> 00:08:47.720 that into better products going forward. 173 00:08:48.760 --> 00:08:51.640 - So in other words, what you do is you combine your 174 00:08:51.640 --> 00:08:55.680 longstanding competitive differentiation 175 00:08:55.680 --> 00:08:59.440 and competitive advantage, which is deep knowledge 176 00:08:59.440 --> 00:09:01.680 of materials, recipes, 177 00:09:01.680 --> 00:09:06.320 and how to deal with them with the technology of the future. 178 00:09:06.320 --> 00:09:09.000 No one can copy the first 179 00:09:09.000 --> 00:09:11.240 and now you make it even better with the second. 180 00:09:11.240 --> 00:09:13.200 - Yeah, but that's exactly right. 181 00:09:13.200 --> 00:09:15.160 But I think that's important for us too. 182 00:09:15.160 --> 00:09:17.040 It's what we call the domain know-how. 183 00:09:18.000 --> 00:09:21.040 That's our know-how that we do have. 184 00:09:21.040 --> 00:09:24.160 And the second example, in our R&D centers 185 00:09:24.160 --> 00:09:26.720 and in our labs, typically you have 186 00:09:27.800 --> 00:09:31.360 testing periods that you have 30, 60, 90 days in order 187 00:09:31.360 --> 00:09:33.200 to see whether the curing is right. 188 00:09:33.200 --> 00:09:36.960 With all the data that we have assembled 189 00:09:36.960 --> 00:09:40.880 in the past, we can actually simulate the outcome 190 00:09:40.880 --> 00:09:44.640 to a very high precision. 191 00:09:44.640 --> 00:09:48.400 So I think that will spare you the physical tests at some. 192 00:09:48.400 --> 00:09:50.400 I'm absolutely convinced that will spare you the physical 193 00:09:50.400 --> 00:09:52.480 tests down the road at some point. 194 00:09:52.480 --> 00:09:53.840 That's huge. We can only do this 195 00:09:53.840 --> 00:09:57.760 because we have the historical data that we can use 196 00:09:57.760 --> 00:09:59.040 to train the algorithms. 197 00:09:59.040 --> 00:10:02.480 So there are many many examples of that in order to use 198 00:10:02.480 --> 00:10:04.760 what we have in our domain as a domain, 199 00:10:04.760 --> 00:10:07.120 Heidelberg Materials Know-How, leverage that in order 200 00:10:07.120 --> 00:10:10.160 to get much better, much faster, 201 00:10:10.160 --> 00:10:12.640 and serve our customers even better. 202 00:10:12.640 --> 00:10:15.800 - And that means you optimize the core value-adding 203 00:10:15.800 --> 00:10:17.920 processes of Heidelberg Materials. 204 00:10:17.920 --> 00:10:19.400 You can do a lot in support processes, 205 00:10:19.400 --> 00:10:21.800 but here you are talking about the core 206 00:10:21.800 --> 00:10:25.400 of your supply chain, of your day-to-day operations, which 207 00:10:25.400 --> 00:10:27.720 coming back to the theme of it needs to tie 208 00:10:27.720 --> 00:10:30.120 to profitability in the end, of course, 209 00:10:30.120 --> 00:10:31.960 impacts a big part of your P&L. 210 00:10:31.960 --> 00:10:34.840 - Absolutely. I think it's, yes, there is opportunities 211 00:10:34.840 --> 00:10:36.400 in the administrative processes, 212 00:10:36.400 --> 00:10:37.920 and I don't want to downplay that. 213 00:10:37.920 --> 00:10:39.360 I think it's clearly there, 214 00:10:39.360 --> 00:10:41.440 but that's, I think that everybody can do. 215 00:10:42.320 --> 00:10:44.240 And differentiate ourselves in 216 00:10:44.240 --> 00:10:46.560 that respect is probably quite difficult. 217 00:10:46.560 --> 00:10:47.880 But if we go to, going back 218 00:10:47.880 --> 00:10:50.160 to our earliest discussion points, if we go 219 00:10:50.160 --> 00:10:52.680 to the core processes of our value chain 220 00:10:52.680 --> 00:10:55.080 and really get going there, A, scale 221 00:10:55.080 --> 00:10:58.200 and differentiation are the two levers that 222 00:10:59.160 --> 00:11:00.880 will then ensure the competitive 223 00:11:00.880 --> 00:11:02.360 edge of Heidelberg Materials. 224 00:11:02.360 --> 00:11:05.200 - Dominik, thank you so much. It's great to see you again. 225 00:11:05.200 --> 00:11:09.000 I hugely appreciate your time and the excitement. 226 00:11:09.000 --> 00:11:10.880 - Thanks so much for having me, Christoph. All the best.