Factories that Think
22 Jan 2026 08:00h - 08:45h
Factories that Think
Session at a glance
Summary
This panel discussion focused on “factories that think,” exploring how artificial intelligence, digital twins, and robotics are transforming modern manufacturing. The conversation featured executives from Siemens, Koch Holdings, Automation Anywhere, and Agility Robotics, along with the UAE’s Minister of Foreign Trade. Roland Busch from Siemens explained their partnership with NVIDIA to create an industrial AI operating system that uses physics-based digital twins to simulate entire manufacturing processes before building them, resulting in 20-25% higher productivity and 20% lower energy costs. The panelists emphasized that digital twin technology has significantly improved due to enhanced computing power and the vast amounts of data that manufacturing has been collecting for years.
Real-world applications are already delivering substantial results, with companies reporting productivity increases of up to 70% and forecasting accuracy reaching 95%. Levent Cakiroglu noted that his company operates six World Economic Forum-recognized lighthouse factories that serve as engines for transferring proven capabilities across their global operations. The UAE has implemented comprehensive digitalization across its logistics network, reducing customs clearance times from days to minutes and applying AI in oil and gas operations to optimize production without traditional shutdowns.
Humanoid robotics represents another frontier, with robots like Digit designed to work in unstructured environments alongside humans. These robots address labor shortages in dangerous, repetitive jobs while requiring strict safety protocols. However, significant challenges remain, including cybersecurity risks, the need for substantial R&D investment, and infrastructure requirements. The panelists agreed that successful implementation requires a holistic approach, starting with clear business outcomes rather than technology promises, and emphasized that cybersecurity must be designed into systems from the beginning rather than added as an afterthought.
Keypoints
Major Discussion Points:
– Digital Twins and Industrial AI Operating Systems: The development of physics-based digital twins that simulate entire manufacturing processes before physical implementation, enabling 20-25% higher productivity and 20% less energy costs. Siemens and NVIDIA are partnering to create an AI-based industrial operating system that makes this technology more accessible.
– Real-World Implementation and Results: Companies are already seeing significant benefits from AI and digital twin deployment, with examples including 70% productivity increases in automotive manufacturing, 95% forecasting accuracy in supply chains, and $200 million savings in inventory management through autonomous AI agents.
– Humanoid Robotics in Manufacturing: The emergence of humanoid robots that can operate in unstructured environments and perform human-like tasks, addressing labor shortages and workplace safety issues. These robots are transitioning from caged, repetitive operations to more flexible, AI-powered systems that can work alongside humans.
– Global Implementation Challenges and Opportunities: The discussion of barriers to technology adoption, particularly in developing regions like Africa, including the need for energy infrastructure, logistics, land accessibility, talent availability, and long-term political stability for successful technology transfer.
– Cybersecurity in Smart Manufacturing: The critical importance of building security into industrial AI systems from the ground up, as increased connectivity and data collection create new vulnerabilities that require end-to-end protection and human awareness training.
Overall Purpose:
The discussion aimed to explore the current state and future potential of AI-powered “factories that think” – examining how digital twins, robotics, and AI agents are transforming manufacturing through real-world case studies, implementation challenges, and the broader implications for global trade and economic development.
Overall Tone:
The tone was consistently optimistic and forward-looking throughout the conversation. Panelists spoke with confidence about current achievements and future possibilities, sharing concrete examples and measurable results. While they acknowledged significant challenges (cybersecurity, implementation barriers, workforce transitions), these were presented as solvable problems rather than insurmountable obstacles. The discussion maintained a collaborative, educational atmosphere with industry leaders sharing practical insights and advice for audience members’ specific situations.
Speakers
– Jamie Heller: Editor-in-Chief, moderator of the panel discussion on “factories that think”
– Roland Busch: President and Chief Executive Officer of Siemens, Germany – expertise in industrial AI operating systems, digital twins, and manufacturing technology
– Levent Cakiroglu: Chief Executive Officer of Koch Holdings, Turkey – expertise in manufacturing operations across diverse businesses including home appliances, automotive, and oil refining
– Mihir Shukla: Chief Executive Officer, Chairman and Co-Founder of Automation Anywhere – expertise in AI agents, automation technology, and agentic AI solutions
– Peggy Johnson: Chief Executive Officer of Agility Robotics – expertise in humanoid robotics, particularly the humanoid robot “Digit”
– Thani Ahmed Al Zeyoudi: Minister of Foreign Trade of the UAE – expertise in government policy on digitalization, logistics, manufacturing, and technology deployment
– Audience: Multiple audience members who asked questions during the panel, including:
– Eniton Tadjoshu from Nigeria (global shaper working on healthcare manufacturing)
– Dipali Goenka from India (leads a home textile manufacturing company)
– Julia from Italy (global shaper working in human-robot collaboration at Italian Institute of Technology)
Additional speakers:
None – all speakers mentioned in the transcript were included in the provided speakers names list.
Full session report
Comprehensive Report: “Factories That Think” Panel Discussion
Executive Summary
This panel discussion, moderated by Jamie Heller, Editor-in-Chief, explored the transformative impact of artificial intelligence, digital twins, and robotics on modern manufacturing. The conversation brought together industry leaders from Siemens, Koch Holdings, Automation Anywhere, and Agility Robotics, alongside the UAE’s Minister of Foreign Trade, revealing that “factories that think” are delivering measurable results across global manufacturing operations.
The discussion demonstrated strong consensus on key technological capabilities, with central themes including the proven effectiveness of digital twin technology, the critical importance of cybersecurity-by-design, and the emergence of humanoid robotics as a breakthrough solution for unstructured manufacturing environments.
Key Technological Developments and Current Capabilities
Digital Twins and AI-Powered Manufacturing
Roland Busch from Siemens presented compelling evidence of digital twin technology’s maturity through the company’s partnership with NVIDIA, which has produced an industrial AI operating system utilizing physics-based digital twins. This approach consistently delivers 20-25% higher productivity and 20% lower energy costs across implementations by simulating entire manufacturing processes before physical deployment.
Mihir Shukla from Automation Anywhere reinforced these findings, noting that manufacturing has been “collecting data a lot longer than many other industries,” providing the foundation for AI-powered operations. He described a European logistics customer managing 685,000 items across 32 warehouses four times daily, achieving $200 million in inventory cost savings. “There is no amount of human workforce that could do that,” Shukla observed.
Levent Cakiroglu from Koch Holdings provided additional validation, reporting 70% productivity increases in automotive manufacturing and 13% improvement in equipment efficiency. His company operates six World Economic Forum-recognized lighthouse factories—advanced manufacturing facilities that serve as innovation hubs—transferring proven capabilities across their global operations.
Real-World Applications and Measurable Results
Minister Thani Ahmed Al Zeyoudi described the UAE’s comprehensive digitization program, which has applied digital twins and AI across the country’s entire logistics network, leveraging connections to 250 ports worldwide. The implementation reduced customs clearance times from days to minutes and enabled oil and gas production improvements without traditional shutdowns.
These implementations have achieved 95% forecasting accuracy in supply chain management, with the technology being applied across entire value chains from extraction to distribution. Al Zeyoudi also highlighted infrastructure projects like Egypt’s 2,000-kilometer railway connecting the Red Sea and Mediterranean, demonstrating large-scale digital integration.
Humanoid Robotics: Addressing Unstructured Environments
Peggy Johnson from Agility Robotics presented humanoid robots as a breakthrough for manufacturing’s most persistent challenges. Unlike traditional caged robots that perform repetitive tasks with precision but lack flexibility, humanoid robots like “Digit”—weighing 200 pounds and capable of lifting 25 kilograms or 50 pounds—can operate in unstructured environments designed for humans.
“We’re not trying to replace humans,” Johnson clarified, “we’re trying to allow humans to do the higher-value work.” These robots address critical labor shortages in dangerous, repetitive jobs while offering operational efficiency by performing multiple tasks throughout the day without extensive reprogramming.
However, significant safety challenges remain. Current humanoid robots operate in work cells and require strict functional safety standards—comprehensive safety protocols ensuring robots can work safely alongside humans—before they can operate without barriers.
Implementation Strategies and Business Considerations
Business-Outcome-Driven Approach
A significant theme emerged around prioritizing business outcomes over technological capabilities. Shukla articulated this philosophy clearly: “I wouldn’t start with technology at all. I would look at, there are only three outcomes you can have. You can either increase revenue, reduce managed cost or reduce risk… chase purpose not promises of technology.”
This approach provides a practical framework for organizations evaluating AI and robotics investments, focusing on measurable financial returns rather than technological sophistication. Busch supported this perspective while emphasizing the importance of beginning with data collection from manufacturing machines and breaking down data silos.
Scaling Challenges and Organizational Requirements
Cakiroglu highlighted that successful scaling requires comprehensive organizational transformation: “Successful scaling requires strengthening governance, operating systems, talent, and organizational culture, not just deploying more technology.” This insight reveals that organizational readiness often represents a greater challenge than technological capability.
The discussion showed that while many companies successfully implement pilot projects, enterprise-wide deployment requires systematic attention to governance structures, talent development, and cultural transformation alongside technological implementation.
Global Development and Innovation Transfer
Infrastructure and Development Approaches
The panelists revealed different perspectives on development strategies. Busch suggested that successful industrialization follows waves: “energy and logistics infrastructure first, then industry, then higher education and healthcare.” However, Al Zeyoudi argued for immediate investment in research and development, warning: “It’s too late to wait for someone else to do it, because you’re going to be out of the game.”
Partnership Requirements and Training Programs
Al Zeyoudi outlined comprehensive requirements for successful technology partnerships, including long-term political stability, energy supply, land accessibility, logistics connectivity, and talent availability. These interconnected factors suggest that successful partnerships require holistic ecosystem development.
Shukla emphasized substantial training programs in technology transfer, citing an example where his company provided free licenses in the region, training 700 women with 525 finding employment within the first week. This demonstrates the importance of human capital development in technology adoption.
Addressing Implementation Barriers
An audience member from Nigeria’s healthcare sector raised concerns about barriers to technology adoption in Africa, including energy infrastructure and talent availability. The responses revealed the importance of customizing solutions to local conditions while building foundational capabilities.
Cakiroglu shared an example of innovation adaptation, describing solar-powered refrigerators developed for rural areas without reliable electricity, demonstrating how technology can be tailored to specific regional needs.
Cybersecurity: A Critical Foundation
The strongest consensus among panelists concerned cybersecurity’s critical importance. Busch emphasized that modern manufacturing systems require end-to-end cybersecurity design from the beginning, as “attack angles are now 360 degrees.” This represents a fundamental shift from traditional approaches where security was added after system development.
Shukla reinforced this perspective, stating that cybersecurity “cannot be afterthought” and advocating for AI systems designed with fixed parameters for critical functions. Johnson addressed cybersecurity from a data protection perspective, emphasizing that sensitive information should remain on customer premises rather than external clouds.
Particularly striking was Busch’s observation that “the most important reason for an attack and a successful one is the human,” highlighting that despite technological sophistication, human factors remain the weakest link in security systems. Al Zeyoudi emphasized the need for government-industry collaboration and continuous employee awareness programs.
Industry-Specific Applications and Audience Insights
The discussion included perspectives from various industries through audience questions. A representative from Italy’s robotics sector and another from India’s textile industry—which deploys 20,000 people and manufactures for 60 countries—highlighted the global scope of manufacturing transformation and the need for solutions that scale across different industrial contexts.
These interactions demonstrated that while the core technologies are universal, their application requires careful consideration of industry-specific requirements and regional conditions.
Practical Recommendations
The discussion yielded several actionable recommendations for organizations considering AI and robotics implementation:
Immediate Steps:
– Begin with data collection from manufacturing machines to enable predictive maintenance
– Target specific high-impact areas such as inventory management and order processing for quick wins
– Break down data silos to create comprehensive data warehouses before implementing sophisticated AI systems
Strategic Considerations:
– Prioritize business outcomes over technological sophistication
– Design cybersecurity into systems from the beginning rather than adding it later
– Invest in comprehensive training programs and organizational change management
For Humanoid Robotics:
– Focus on achieving functional safety standards as the critical milestone for human-robot collaboration
– Start with controlled environments before expanding to more complex applications
Implications and Future Outlook
The discussion revealed that AI-powered manufacturing represents a fundamental transformation in capability rather than incremental improvement. The scale of operations now possible—such as real-time management of hundreds of thousands of items across multiple locations—creates new possibilities for efficiency and responsiveness.
However, successful implementation requires comprehensive attention to cybersecurity, organizational readiness, and workforce development alongside technological deployment. The consensus among industry leaders suggests that while the technology is ready, successful adoption depends on holistic approaches addressing technical, human, and organizational factors simultaneously.
Conclusion
This panel discussion demonstrated that AI-powered manufacturing has moved from experimental to operational reality, with proven results across multiple industries and geographies. The convergence of digital twins, AI agents, and humanoid robotics is creating manufacturing capabilities that were impossible just a few years ago.
Success requires moving beyond technology-first thinking to business-outcome-driven approaches, with cybersecurity designed into systems from the beginning and comprehensive attention to organizational readiness. The alignment among diverse stakeholders suggests that best practices are emerging, creating opportunities for scaling these technologies globally while addressing the unique challenges and requirements of different regions and industries.
Session transcript
Good morning. Thank you so much for being here this morning. We’re talking about factories that think.
I’m Jamie Heller. I’m the Editor-in-Chief of … panel, Roland Busch, President and Chief Executive Officer of Siemens, Germany, Levent Cakiroglu, Chief Executive Officer of Koch Holdings, from Turkey.
We’ve got Mihir Shukla, Chief Executive Officer, Chairman and Co-Founder of Automation Anywhere, and Peggy Johnson, Chief Executive Officer of Agility Robotics. So thank you all for being here, and thanks to the audience for being here. Roland, I want to start with you.
Siemens and NVIDIA have recently announced a partnership to build an industrial AI operating system. Could you tell us what that is and why it’s important?
Okay. I’ll try to make it short. So let’s assume now you want to build a new manufacturing line for electronics.
What you do is you simulate the whole thing. You make a digital twin of the product you want to manufacture, of the production line which you’re going to use, so of the whole environment, before you even start building it. That allows you to optimize the product, the manufacturing line, in a very comprehensive way.
This is where it starts. And here we go. So this is used digital twins.
And when we talk digital twins, we talk really physics-based digital twins. So they behave like the real thing when you heat it, when you run power on it. And then when you run software on it, we can simulate how software runs on silicon.
So once you have it, then you start sending an excavation machine, you build your plant, and you run it. Output is 20, 25% higher space productivity, 20% less energy cost. You’re much, much faster in ramping up.
You don’t make mistakes.
If you use a digital twin.
Exactly.
You see these things.
Then comes, and you want to have this digital twin alive once you are running your production because it allows you to optimize all the time. If you make a change, if there’s something going wrong. You see real-time what’s happening on the shop floor because you have a digital twin which is still alive.
It’s feeding by the real-time data and they can optimize. If a new product comes in, a new component, you know exactly what’s happening. And for that you need a technology stack, you need data, you need a very solid software, simulation software.
You need an operation software which runs your manufacturing site and applications which help you operate. This is what we call the operating system, the industrial operating system, AI-based. And you need a lot of technology.
This is where we are partnering with NVIDIA. They have Omniverse. They can make a photorealistic view of this whole manufacturing line.
So you look at it as if it would be a video and you can really simulate everything, including, by the way, the people working on the shop floor. So it’s not only the machines but also the people. And that’s very powerful.
And what we want to do is we want to make this operating system as easy as possible to use for our customers so they can scale it and use it in a greenfield application, that’s what I explained, or brownfield.
You don’t have to start from nowhere. You can start at any point in time.
So the operating system is going to be a product that you’re going to sell to your customers?
If you consider a technology stack as a product, yes, it is. So a product means that you cannot plug and play it for each and every application. You really have to adapt it to your individual needs, obviously.
It makes a difference whether you manufacture a car, you run a semiconductor plant, or you make a yogurt or a pharmaceutical. It works in all cases, but it needs, obviously, specialties. They differ.
But the principle that I explained is always the same. Whether you build a product or otherwise you simulate a molecule which finally runs in a drug manufacturing plant, the whole model is the same.
Got it. Okay. I want to come back to this, but first I want to introduce one more participant, the Minister of Foreign Trade of the UAE, Thani Ahmed Al Zeyoudi.
Thank you, thank you for being here Let’s get right into it. We’re talking about digital twins and AI systems. Oh, you’re still waiting or sitting?
Okay, so Mihir, can you just bring us up to speed on the digital twins? Like they’ve been around but they’re getting better. How how important are they?
What’s what’s exciting about them?
You’re absolutely correct. The digital twins have been around but the technology has exponentially improved and the computing power is exponentially improved. One of the unique advantage manufacturing has is that they have been collecting data from every machine and every system a lot longer than any many other industries.
So there’s an enormous amount of information feeding all of those information with new more powerful models allow you to create what Roland was talking about a digital twin that exhibits all the characteristics of the physical system.
The and with the power of AI agents, this is not I’m gonna expand the theme from not factories that think but they think they connect and act, right? It’s a it’s a it is proactive because let’s say digital twins tell you that there is going to be a problem in manufacturing. You can proactively act on it.
It has implications on your inventory. It has implications on your shipping planning and so can it can it alert everybody and do countermeasures across the sub because that’s where the costs are, right? So, how can it proactively operate with AI agents can can can instantly react and bring that resilience across entire entire supply chain and operations and factories are at the heart of it, right?
And so if you if you catch something there you can make everything better.
Well, then you have factories you have plants all over the world. What we’re hearing how how real is it right now or how much of it is something that’s looking to the future?
We are operating with quite diverse businesses and digitization is a group-wide imperative for us. We scale what works across home appliances, automotive, and oil refining. Currently, we are operating with six WEF-recognized lighthouses.
In addition to value creation in terms of quality, productivity, efficiency, and sustainability, these factories are engines for us to transfer already proven capabilities. In that respect, what we have been talking about are the ones that we have already applied and benefited from. Take digital twins.
Digital twins are emerging as productivity and flexibility engines. In the world that we are operating, we have to be operating in real, live data and decision tools. In that respect, in manufacturing, digital twins technology enables us to increase productivity and flexibility.
At our car manufacturing plant, the productivity increased by 70% and overall equipment efficiency increased by 13%. In oil refining, by using digital twins and integrating yield, energy, and blending decisions together, we have been able to increase the capacity utilization. In supply chain and supply chain management, combining planning and financial decisions, We have been able to increase the forecasting accuracy up to 95%.
So it is real, it is working.
It’s happening now.
It’s happening now.
Are you ready to go? Minister, how impactful is this for global markets and global trade? Is it still just a regional phenomenon or what are you seeing?
First of all, the digitalizations and the digital twinning, as we’ve been saying, is coming along. And it’s coming not only in factories, but it comes on all value chain of the productions lines, from the extractions to the manufacturing to the logistics to the distributions, including the customs clearances, et cetera.
We at the UAE, we have a very interesting experience because we’ve been investing heavily on logistics and our global network. We’re connected to 250 ports around the world. And we’re continuing growing this network because it’s critical.
Now, what we did at Back Home Logistics, we start transforming all of our operations in the customs and ports to be digital. The clearances for the majority of the consignments happens before they arrive. The payments or the deposits, returns, it’s an immediate action.
They don’t need to wait for days. The clearances used to take a few days. Now it’s just a matter of a few minutes and they’re done, et cetera.
After the sophistication that we have done internally, now we’re taking this experience globally. And wherever we manage, we do that. And we associate most of the industrial zones close to the ports where we manage because that is going to make things much faster and movements of the commodities that we produce there.
Now, what we have done in the factories or the manufacturers. sector in the country. Eight years ago, we started having a dedicated technology aspect within the Ministry of Manufacturing.
And we were sending a message that technology has to take over the huge manpower that we’re bringing from abroad. It’s not anymore about the wages, it’s about digitalization, it’s about improving the efficiency of how we’re running the operation. How we have done it, it’s transformative.
We’re applying it, we’re deploying technology, and we’re moving forward. Some of the examples in oil and gas. Oil and gas, for sure, you produce something which is far away, a kilometer at least below the ground.
So the simulation assisted us in improving the efficiency of the productions, producing the right wells and controlling the pressure of the wells. Now, moving forward, we’re applying the AI. The AI tells us where to produce, et cetera.
We don’t need anymore the simulation engineers to tell us where. The AI is giving us. And we have already built that sophisticated system.
Now, we want even to the operation, we start applying the technology, the robotics, and the sites where we used to have a shutdown for three to six months. Now, by robotics, we’re just applying it while doing the operation. He’s doing the robots.
He’s doing the surveillance. The situation and apply it without having the shutdown that used to take months. Now, it’s just a matter of a couple of days and the operation is back.
Moving forward as well to some of the heavy industry. Aluminum, for example. We start applying the AI and smelting refineries.
And that is improving the efficiency, the production, et cetera. So, for sure, the twin and digitalizations have a huge impact. It’s just a matter of the well that you want to do it.
And you have to do it. Because it is coming. And with the challenges and the economic status facing the world.
Without digitalization, we will not be able to reduce the consequences of those measures.
Okay, so you need the will and Peggy, I want to get you in here because you have a humanoid robot named Digit. Yes. And robotics has been around a long time.
It’s just starting to make some major breakthroughs, but it’s hard, right? I mean, it sounds really good, but can you talk about some of the challenges of doing this safely and…
Sure, sure. So humanoid robots are highly complex. They’re a little bit like EV vehicles.
They have to take in a lot of data. They have to process it quickly. They have to make decisions, which is why we’re just starting to see them.
Clearly, AI has helped supercharge it a bit, but you still need a dynamically stable big device that can add value to a factory. So it’s been it’s been slow in coming, but I would say that one of the turnarounds recently with humanoids is that they can operate in unstructured environments. Many robotics are…
Unstructured? Unstructured. That’s the most recent development.
Yeah, and what I mean by that is they can step into places that are built for humans and do human work. It’s largely material handling, now moving bins and boxes and things like that, item manipulation. But that’s in contrast to sort of fixed robotics, maybe conveyor belts or single arm things.
Those are obviously very functional and produce a lot of output. But in a humanoid, you can have something that has a lot of operational efficiency because it can do one job in the morning and something else in the afternoon, and it can go where humans go. It can go down narrow aisles.
The world is built for humans, and so it can reach up high. It can go down to the floor, and that puts it in a category different from some other robotics, which are really meant to be static and in one place and do one job very well. So now that they can operate in these unstructured environments and with the advent of AI we can kind of supercharge the training of robots and Get them out there and to do more and more things and as their capabilities grow They’ll start to move beyond enterprise and eventually into our homes, but you mentioned safety.
That’s a factor They have to be meet a very high safety bar to go from a factory into our homes Because right now in factories, they’re mostly in like cells right in like correct yes, they are meant today all humanoids have to operate inside of a work cell and You might see videos and demos and you know making coffee and doing backflips and things But when you actually put them to work, they’re automated machinery and they’re 200 pounds They can lift 25 kilograms or 50 pounds like these things are powerful so they must stay inside of a work cell away from humans and The bar that you have to meet to get outside of the work cell is called being functionally safe So you want to be able to be in close proximity with humans?
But to do that you have to not harm the humans when you get close to them So it has to sense that I my robot has to sense It’s approaching a human and as it gets closer and closer the radius gets smaller It’ll bring itself down to the ground and it’ll keep carrying its payload.
It has to do all that balancing When the human passes by it’ll stand up again and move on its way We’ll have one of those by the end of this year a functionally safe robot. So then it can run around the factory.
Just just as a part that man But you what you are talking about is the caged robots, when you typically when you see a manufacturing line for cars the welding robots caged and they do a repetitive task as precise as possible and they do it extremely fast, but it’s always the same.
It’s always the same just welding the same parts over and over again. This environment is completely different And so you do don’t need really too much sensing Security is defined by the cage in that case It’s a completely different environment the input tokens are sensors different environment and robots act accordingly that works only with AI Technologies correct yeah, and that’s that’s what what you’re talking about and it starts really in logistics Where you see a lot of handling stuff, but that that moves more and more into different areas of manufacturing
And and the problem we’re trying to solve is the job some of the jobs that people don’t want to do is … There’s a couple things one is it’s very hard to hire for these manual jobs. They’re dull and dirty and they’re dangerous at times because you’re lifting over and over again very repetitive It’s kind of mind-numbing work Typically people turn over within a year Very quickly they you know they’re they’re they’re moving along But the injuries are also another thing we have an aging Workforced a lot of young kids don’t want to go into environments like this so, the older employees are also getting hurt more.
They’ve been doing this manual work for longer. There’s high costs there And so those two things are really driving the need for humanoids because it’s hard It’s so hard to find humans to step into these roles and really humans shouldn’t you know lift things over and over again and hurt Their knees and backs.
There’s other jobs. There’s higher order value that humans can offer versus just moving. You know item manipulation and moving boxes.
Minister I want to come back to you. You said there’s so much potential if you have the will What what what do you mean? Is it money?
Is it political will is it? Public sentiment open to this tell it what is the challenge?
See the the most important thing is the you know the issue the challenge and you work toward it sometimes you have the money, but you you haven’t diagnosis the right solution to the problem that you’re having I’ll give you an example, as a nation, we welcome everyone and we are attracting the talents.
And one of the challenges that we have, we’re heavily dependent on the non-skilled labours for the constructions, which is exactly what my colleague was talking about. It’s becoming ageing, it’s too difficult to attract new young people to start working on the constructions, and robotics is the main solution. Are we going to stop there?
Absolutely not. The construction is part of the developments of the nation, and we’re going to have, we’re having the largest development projects in the region. So we’re heavily investing in robotics factories to ensure that it does this job, instead of continuing bringing or looking for such labours.
Finance is critical, for sure, but many countries are having the finance, but they don’t have the right solutions or the right diagnosis to sort out the issues. What we have done in the country, we came up with a very holistic approach when it comes to robotics and AI. Starting with the government appointments, the policy developments, capacity buildings, but the main game changer that we have done is the R&D investments, which is usually a very costly matter, and many stakeholders are avoiding that because they say someone else will do it, I’ll just copy paste it.
But when it comes to AI, it’s too late, and digitalization technology, it’s too late to wait for someone else to do it, because you’re going to be out of the game.
Okay, just hire Siemens?
We’re working with them. But you have to start working on R&D, because you ensure to customise things to your own ecosystem, your own environments, your own conditions, et cetera, and then you start deploying and taking things forward.
The finance is combined with that, because you need to have the entrepreneur supported to bring those ideas and R&D thoughts into practicalities and commercial phase, which we’re doing through different stakeholders.
We’re applying in the country the two models, the deployments and investments model, where we invest in international technology and bring it to the country, but also we have the consumer market where we apply anyone’s technology in the country who would like to pilot and experience their projects.
We’re becoming a regulatory lab to everyone, and at the same time we’re doing our own investments, which match our own interests, et cetera. So coming back to your question. The will is there and should be there, but you have to know what you want, and you have to wait for a long term.
You cannot just do it overnight and think it’s going to happen, because if you don’t have that patience and consistency, you will not be able to achieve whatever you want.
Levent, your company, how much are you doing your own investments, your own development versus working with partners and other providers?
Of course, we need to work and collaborate with technology partners, but in addition to that, we have been building up our capacities, capabilities in digital space. In that respect, I can give an example of Platform 360 that we have developed in-house, and it has been deployed in more than 40 sites. It scales execution and decision quality across the businesses.
As complexity increases, the need to speed, accelerate the response speed and consistency become a real constraint. In that respect, this platform standardizes the diagnosis and decision across our businesses, and it is one of the most important aspects from our point of view to scale from pilot to the enterprise level.
Is the ability to evaluate and make these decisions, that’s the technology? What is the technology you’re saying that is so important to scale across?
We have been deploying artificial intelligence, machine learning, IoT, all the digital twins technologies that we have talked about. But it’s important to… have a single architecture to utilize what we have as a digital capacity across the businesses.
One of the aspects of intelligent factories is to be able to scale what is achieved in pilot studies and it is not about deploying more technologies. It is more about strengthening the governance, operating systems, talent aspect and the culture and leadership within the organization to make sure that we benefit from our capabilities across the world.
I want to go to the audience for questions, but first Mihir, I just want to come to you which is, we are talking about R&D and the importance and trying new things and yet companies have a business to run, they have numbers to make and how do they, what are you hearing from your clients, like what’s, they probably feel a lot of pressure to get with the program and get going, but they also have their daily obligations and also is it better to be starting anew with fresh technology or just trying to like AI-ify what you already have?
So I think that the, we have maybe thousands of customers in manufacturing vertical alone, running AI agents in millions, the possibilities and the benefits are enormous and some of them are within three months, so you can unlock a vast amount of capital.
Let me give some examples. Take an example of Sumitomo Rubbers, they started using this in the planning and then in inventory management and eventually as they started connecting AI agents, one of their biggest benefit came in container planning and shipping.
Because if you optimize that effectively, there was a huge dollar amount. Take another customer, Cargill, just in order management and order to cash processes. You can unlock, let’s say if you do a straight through order processing, increase that percentage by 20 or 30 percent more, you can unlock a vast amount of revenue stream.
So by targeting these specific areas, I’ll pick one more on inventory management. We have a European customer, and we were trying to see how to create an autonomous inventory management. And initially, we were trying to mimic what human beings were doing, and then it occurred to us, why are we doing it?
This is a lot more powerful. So we created AI agents that dynamically manage inventory in 32 warehouses and 685,000 items, four times a day. There is no amount of human workforce that could do that.
And it saved them $200 million in inventory costs. So each of these things could unlock a vast amount of capital that is being vested in this current system. You take all of that, you take the will, you take the multi-year planning, and reinvest it into a sustained competitive advantage.
This is a phenomenal time to be doing this.
So it’s really not optional, and you could just keep getting just better margins to simple stuff. Do we have any questions in the room here? Over here, please just let us know, if you don’t mind standing up, your name, where you’re from, and your question.
Thank you.
Hello, everyone. My name is Eniton Tadjoshu. I’m a global shaper from Nigeria.
And I work for Nigeria’s presidential initiative for unlocking the healthcare value chain. So that’s trying to boost local healthcare manufacturing. One of the ways that we do that is through technology transfer.
And so, my question is, one, I want to learn more about the work you all are doing in Africa, not just as a market to sell to, but as a partner to build industry, and also to know what defines a great partner when it comes to technology transfer.
What are you looking for when you want to look towards sustainable growth of industry in Africa, specifically?
Do you want to take that, or?
I can give it a try. Maybe I give one example, and what we are doing in Africa, big time is, and it starts in northern Africa, in Egypt. Here we have the largest project actually Siemens ever did in the history, which is about 2,000 kilometers of railway lines, from the rail to the signaling system, electrification, and the trains, the rolling stock, the stations, connecting Red Sea and Mediterranean Sea, connecting the north and the south down to Sudan.
So, and this changes the whole economy of this system, because we’re connecting 90 million people. And there’s a lot of technology which is going in. This includes, of course, by nature, technology transfer, because finally you operate the system on the ground.
You have people how to know how to make a kind of a re-overhaul of your trains, run signaling systems. Actually, it’s European ETCS technology which comes to the country. So this, along with that project, it’s not only boosting the economy, because it obviously changes the logistic systems completely, the way how you connect people, including obviously, call it tourist line, which is connecting Rwanda and Luxor.
But it really brings technology to the people. We do that, by the way, in training people. Siemens invests 430 million just training our own people.
on new technology, AI technology. We do that every year, but we are opening this platform also to our customers and our partners to take advantage of the technology which we have because it requires a lot of knowledge. Maybe one word on Africa.
This is not about AI technology, this is a general one. And when economies develop, they develop in four ways. Energy, logistics infrastructure, so communication, mobility, then comes industry, and then comes higher education, healthcare.
It goes concurrently, but you see these waves. And not before you have the first two. You really are able to build up substantial industrial technology.
And that’s maybe one of the biggest problems of many, many countries in Africa that they are in the wave one and two. Egypt is moving straight forward to obviously the next one. So, and then last point here, the advantages now, what I encourage is, if you’re anyhow building that up now in that world we are living in, just try to leapfrog and use AI technologies all the way from the beginning.
It’s much, much easier if you do that right away from the beginning, rather than have a brownfield where you start migrating. Keep it short, it’s a long story.
And you mentioned talent and training. One of the things we were able to do is give free licenses and capabilities to the region. In one particular case, we trained 700 women.
And the great part is that they don’t have to unlearn and this technology is so natural to learn that it doesn’t take four year degree to learn them. So, there is an instant mobility. And out of that 700 women, 525 found a job in first week.
This is remarkable, right? And this is agentic AI jobs. So, I think there are multiple dimension to it, but talent is very important.
We have operations in Egypt and in South Africa, factories. And in addition to the products with the highest energy and water efficiency, we also transfer the capabilities that I have mentioned here to those factories as well. Those definitely create value in the countries that we operate in.
And that includes investment in our talent in order to make sure that the technologies are being actively used and further developed by our colleagues. Maybe out of intelligent factories, I can give an example which is close to my heart. Our colleagues have developed a refrigerator which is powered by solar.
It is quite vital in rural areas where the access to electricity is very limited.
And a refrigerator powered by solar. You better have an open roof.
If you allow me to jump in here because, you know, South Africa, Egypt are more well structured when it comes to investments. But we start as well as a nation investing in those LDCs within Africa. But the answers will not be only applied to Africa, but everywhere.
There are a few things which any partnership has to take into consideration. The long term, there’s no disturbance on this investment. That’s a critical matter.
Second one is energy, which you mentioned. You have to ensure that there is enough supply of energy to power those manufacturers, technology supply areas, et cetera. Third one, availability of land.
Sometimes lands are not owned by governments, owned by people. So it’s going to be very challenging to have access to the land to do those factories. Land accessibility.
And the fourth one, logistics, which is even if you do the investments, you have to ensure that there is a connectivity between wherever you are to the poorest or to the closest area. Fifth one, which is as well I see it as an important one, to guarantee at least the movements of the project is the off taker. Is it going to be the government guaranteeing that or the private sector?
And the last one is availability of talent, which is an issue nowadays because everyone is trying to pull the talents when it comes to technology to their part. So it’s an ecosystem. You cannot just take one part of it and keep the others.
If you don’t have all of them, you don’t move on. Some of the countries outside Africa as well, they have the lands and they give us the accessibility to power. but the logistics is a nightmare, so you cannot move on on that.
So even if you have the main ones, you have to look at the whole thing because otherwise the visibility of the projects will not be there.
That’s a lot of challenges. Great question, thank you so much. Do we have another question in the room?
And behind you as well. Thank you so much. Please stand up.
Hi, I’m Dipali Goenka from India and I lead a home textile manufacturing company which is one of the largest in the world and I manufacture for 60 countries across the world. My supply chain is very complex. It starts from the farmer to spinning to weaving to processing to cut and sew.
My question to the panel here is when do you start looking at digital twin and do you look at it in a holistic perspective because I deploy 20,000 people. I’ve actually put in, you know, data, you know, where machines can speak to one another. How do you use it and how do you look at using digital twin?
And when you talk about, you know, business, you’re talking about return on capital employed as well. How do you look at it? So I think for me the question is when do you start deploying digital twin in a complex supply chain where your customers are sitting in United States of America, UK, Europe, Japan or Australia and you are looking at end-to-end.
I’ll start. I run a technology company and I would tell you I wouldn’t start with technology at all. Is digital twin the right answers?
Maybe not. I would look at, there are only three outcomes you can have. You can either increase revenue, reduce managed cost or reduce risk, right?
And so in your business, where is the biggest benefit? Maybe part of running a resilient supply chain, you can reduce cost effectively and generate $50 million of benefit. I would do that first and use that money to fund everything else, create more oxygen further.
So I would prioritize based on business outcome. We often tell customers that chase purpose not promises of technology, and I don’t know technology company But that’s otherwise you will always have a gap between AI’s promise and impact of AI and the diffusion of AI Technology is a means to an end Amazing technology by the way, but means to it
Do you have a let me add? I mean I know a little bit of textile industry. But but everything and you rephrased it yes, it’s it starts with your supply obviously then you then you manufacture.
I assume you have a lot of manufacturing machines. Which are which are complicated, and you want to let them run 24-7 obviously and and then and then comes the the demands. Let me for the in the middle part and if it comes to machines.
This is where I believe you could really start Working on and you don’t need a digital twin of your of your textile machines. But what you need in the first place is collecting the data how they operate so start with the data layer. So you collect the data, and then you can and this is the beauty about it once you have a good call it operating software which controls your machines.
They can show AI on it, and they tell you this machine is going to fail in a week from now. So you and they tell even for that reason so you can order a spare part this machine doesn’t stay a week. It stays maybe just an hour because you know exactly what’s happening.
This is possible with AI technology right now and and again and this this makes your Capex run 24-7 makes a big difference. I’m even not talking about getting more in autonomous operations, which is in some cases easy and difficult. I mean pretty sure you have your suppliers for your machines And that’s something where I mean we have this technology.
We could could could support also to do that that piece the other one is the supply chain if you really want to see how, how do you get this whole thing running? Where can you cost? Can you cost out then you’re more in the ERP system?
Which you which you hopefully obviously run as well and the best is if you start tearing down data silos. I’m pretty sure you have data silos in your supply, in your machines, in your cell. Start tearing these silos down, create a warehouse, data lake, whatever.
Then you can, once you have that, you can do exactly what you said. You can connect the data, AI helps you. And this is super powerful because it opens you, it’s a view on your operations which you never had before.
I mean, I might have just paid for your ticket to Switzerland, some great advice there. We have one more question, I think, back here.
My name is Julia. I’m also a global shaper from Italy this time. I work in human-robot collaboration at Italian Institute of Technology.
My question is that, I mean, technologies are fundamental, bringing them to the factories but I think that increases, it’s at their exposition to cyber security risks. So, just imagine if malware stops production, that’s a great damage, right? So, I was wondering how are you addressing this problem, if you are?
Thank you.
You wanna get in here, Peggy, because these robots are pretty risky.
They are. When you think about what your mobile phone can do with the sensors on that, this is a different level because of the sensors that are on robots. Obviously, they have LIDAR, they have cameras, they have hearing devices, they are machines inputting data.
It’s a lot of data that’s very, very sensitive. So, from the very beginning of our development of the robots, we knew that that was going to be an issue for any customer. So, we keep as much data on board the customer or offloaded locally to the customers.
sites inside their factory and we don’t own any of the data. We, with their permission, we can help process it for them, for instance, off in the cloud if the robot needs to make a decision or something, but the data is the customers and that was sort of the principle we went in with so that every, the way we built the robot was to keep all of that data protected, to keep the comm links protected and ensured that, you know, others, bad actors couldn’t get to it because clearly that’s very competitive data as well.
I would also, I would also say that as part of design, the, you want to design it so that where do you use artificial intelligence and where there are dynamic reasoning and where you lock it in so that it does not deviate.
I’ll use an extreme example of a, let’s say there’s a cruise missile and it can dynamically decide many things, it’s kind of a robot, but it has a fixed destination. That does not change, that’s not dynamic because you knew it before you launch where it is going, yeah. There are some things you do not want it to be dynamic and in this exuberance of AI, sometimes we replace natural intelligence with artificial stupidity and that’s a wrong answer.
Did you want to pop in here?
Yeah, so I mean the, before before digital technologies, robots and whatever came in, very often you see customer, our customers running a manufacturing site disconnected, disconnected from the rest of the world except one router which is, which we throw every cyber security on it, what you, what you have, so it’s an island.
That doesn’t work anymore. I mean it goes end to end, you start somewhere and a sensor, a thermostat which you can hack eventually all the way to the cloud. So you, the attack angles are now 360 degree, much, much more than before.
So you have to start from the design, from the, from the beginning how you design your cyber security landscape. And what we do is we deploy technology, which is snorkelling on each of the bus systems which we run in a manufacturing site, looking for patterns which are unusual. This goes all the way to the cloud with all the cloud services which you get from your hyperscalers, from your suppliers.
So it has to be an end-to-end kind of solution, and each bit and piece has to come together. Most importantly is you cannot avoid somebody attacking you, you cannot even avoid somebody going behind your firewalls, but what you can avoid is once they are behind that you stop this thing, because you detect something.
And this is, it takes normally, I mean, minutes or maybe half an hour, hours before really damage can happen, and you stop it before that. So there’s much more to talk about it, but it’s a super complex and slow, and we have a huge ecosystem of partners, new technologies, start-ups, they come up with crazy ideas and huge technologies. We are very much tuned to do that and use it as much as we can.
I think you almost have a redesigned security first. Exactly. It can be afterthought, it cannot be afterthought, this is, the question is so important.
If you allow me to jump in from the government point of view, we started early on cyber security, and the other thing which we did, we built the right protocol, and we start making sure that people are aware, you cannot just work in isolation as a cyber security authority, you have to know, to make sure that the employees, anyone who is in factories, or in the government’s employees, they are aware of the protocol and what are the steps that has to be taken.
And for sure, the protection measures, et cetera. But the most important thing is to start from the beginning. Cyber security has to start hand-in-hand with the whole development, otherwise it’s going to be too late.
And ensure that the awareness is going and the development is continuing, it does not stop somewhere.
And the most important reason for an attack and a sucssefull one is the human.
That’s where the greatest risk is, and that’s why it’s so important what you’re saying, to have well-educated humans on this. Lots of risk, but huge potential. Thank you all for an amazing panel.
Thank you for being here. Have a great day. Thank you That’s great, thank you all
Roland Busch
Speech speed
180 words per minute
Speech length
1848 words
Speech time
613 seconds
Digital twins enable comprehensive simulation and optimization of manufacturing processes before and during production, resulting in 20-25% higher productivity and 20% less energy cost
Explanation
Busch explains that digital twins allow companies to simulate entire manufacturing lines and products before building them, creating physics-based models that behave like real systems. This enables comprehensive optimization of both products and production processes, leading to significant efficiency gains.
Evidence
20-25% higher space productivity, 20% less energy cost, much faster ramp-up times, and fewer mistakes. Example of simulating electronics manufacturing lines including the whole environment and even people working on the shop floor.
Major discussion point
Digital Twins and AI-Powered Manufacturing Systems
Topics
Economic | Infrastructure
Agreed with
– Levent Cakiroglu
– Mihir Shukla
Agreed on
Digital twins are delivering real, measurable results in manufacturing
Traditional caged robots perform repetitive tasks precisely but lack the flexibility that AI-powered humanoid robots can provide
Explanation
Busch distinguishes between traditional manufacturing robots that are caged and perform identical repetitive tasks versus new AI-powered robots that can adapt to different environments. He emphasizes that traditional robots rely on cages for security and perform the same welding tasks repeatedly, while AI-enabled robots can sense and adapt to their environment.
Evidence
Example of welding robots in car manufacturing lines that are caged and do repetitive welding of the same parts over and over again, contrasted with AI-powered robots that use sensors and can act accordingly in different environments.
Major discussion point
Humanoid Robotics and Automation Challenges
Topics
Economic | Infrastructure
Economic development follows waves: energy and logistics infrastructure first, then industry, then higher education and healthcare
Explanation
Busch argues that economies develop in a predictable sequence, starting with foundational infrastructure like energy and logistics, followed by industrial development, and finally advancing to higher education and healthcare. He suggests this understanding is crucial for technology deployment strategies in developing regions.
Evidence
Example of Egypt’s railway project – 2,000 kilometers connecting Red Sea and Mediterranean Sea, connecting 90 million people, including technology transfer for train maintenance and European ETCS signaling technology.
Major discussion point
Global Development and Technology Transfer
Topics
Development | Infrastructure
Disagreed with
– Thani Ahmed Al Zeyoudi
Disagreed on
Technology transfer and development strategy
Companies should start with data collection and breaking down data silos before implementing comprehensive digital twin systems
Explanation
Busch advises that companies should begin their digital transformation by collecting machine data and integrating disparate data systems rather than jumping directly to complex digital twin implementations. He emphasizes that once you have integrated data, AI can predict machine failures and optimize operations.
Evidence
Example of textile manufacturing machines – collecting operational data to predict machine failures a week in advance, allowing for spare parts ordering and minimizing downtime to just an hour instead of a week.
Major discussion point
Implementation Strategies and Business Impact
Topics
Economic | Infrastructure
Agreed with
– Mihir Shukla
– Levent Cakiroglu
Agreed on
Business outcomes should drive technology adoption decisions
Disagreed with
– Mihir Shukla
Disagreed on
Implementation approach for digital transformation
Modern manufacturing systems require end-to-end cybersecurity design from the beginning, as attack angles are now 360 degrees
Explanation
Busch explains that unlike traditional isolated manufacturing systems, modern connected systems create multiple attack vectors from sensors to cloud services. He advocates for comprehensive security architecture that can detect and stop attacks quickly, even if initial penetration occurs.
Evidence
Technology that monitors bus systems in manufacturing sites looking for unusual patterns, extending from sensors to cloud services. Detection and response systems that can stop attacks within minutes or hours before significant damage occurs.
Major discussion point
Cybersecurity and Risk Management
Topics
Cybersecurity | Infrastructure
Agreed with
– Mihir Shukla
– Thani Ahmed Al Zeyoudi
– Peggy Johnson
Agreed on
Cybersecurity must be designed from the beginning, not added as an afterthought
Mihir Shukla
Speech speed
160 words per minute
Speech length
934 words
Speech time
348 seconds
Digital twins combined with AI agents create proactive systems that can predict problems and trigger countermeasures across entire supply chains
Explanation
Shukla expands on digital twins by emphasizing their integration with AI agents that not only think but also connect and act proactively. He argues that when digital twins predict manufacturing problems, AI agents can automatically alert relevant parties and implement countermeasures across inventory, shipping, and supply chain operations.
Evidence
Manufacturing has been collecting machine and system data longer than many other industries, providing enormous amounts of information to feed into more powerful AI models that can exhibit all characteristics of physical systems.
Major discussion point
Digital Twins and AI-Powered Manufacturing Systems
Topics
Economic | Infrastructure
Agreed with
– Roland Busch
– Levent Cakiroglu
Agreed on
Digital twins are delivering real, measurable results in manufacturing
Companies should prioritize based on business outcomes (revenue increase, cost reduction, risk reduction) rather than chasing technology promises
Explanation
Shukla advocates for a business-first approach to technology adoption, arguing that companies should identify where they can achieve the biggest financial benefits first and use those gains to fund further technological advancement. He warns against getting caught up in technology promises without clear business purpose.
Evidence
Advice to chase purpose not promises of technology, emphasizing the gap between AI’s promise and actual impact. Recommendation to prioritize based on three outcomes: increase revenue, reduce managed cost, or reduce risk.
Major discussion point
Implementation Strategies and Business Impact
Topics
Economic
Agreed with
– Roland Busch
– Levent Cakiroglu
Agreed on
Business outcomes should drive technology adoption decisions
Disagreed with
– Roland Busch
Disagreed on
Implementation approach for digital transformation
AI agents can unlock significant capital quickly, with examples showing $200 million in inventory cost savings and dramatic improvements in order processing
Explanation
Shukla provides concrete examples of how AI implementation can deliver substantial financial returns within short timeframes. He demonstrates that AI agents can manage complex operations at scales impossible for human workers, leading to massive cost savings and efficiency improvements.
Evidence
Sumitomo Rubbers benefits in container planning and shipping; Cargill’s 20-30% increase in straight-through order processing; European customer saving $200 million through AI agents managing inventory in 32 warehouses with 685,000 items, four times daily.
Major discussion point
Implementation Strategies and Business Impact
Topics
Economic
Technology transfer should include substantial training programs, with examples of 700 women trained resulting in 525 finding jobs within the first week
Explanation
Shukla emphasizes the importance of inclusive technology training programs that can provide immediate economic opportunities. He highlights that AI technology is intuitive enough that it doesn’t require extensive formal education, enabling rapid skill development and job placement.
Evidence
Free licenses provided to regions, training of 700 women with 525 finding agentic AI jobs in the first week, noting that the technology is natural to learn and doesn’t require a four-year degree.
Major discussion point
Global Development and Technology Transfer
Topics
Development | Economic
AI systems should be designed with fixed parameters for critical functions while allowing dynamic reasoning only where appropriate
Explanation
Shukla argues for careful design of AI systems that distinguishes between functions that should remain fixed and those that can be dynamic. He warns against replacing human intelligence with poorly designed AI systems and advocates for intentional limitations in AI decision-making for safety-critical applications.
Evidence
Extreme example of a cruise missile that can make many dynamic decisions but has a fixed, unchangeable destination that was determined before launch.
Major discussion point
Cybersecurity and Risk Management
Topics
Cybersecurity
Agreed with
– Roland Busch
– Thani Ahmed Al Zeyoudi
– Peggy Johnson
Agreed on
Cybersecurity must be designed from the beginning, not added as an afterthought
Levent Cakiroglu
Speech speed
92 words per minute
Speech length
526 words
Speech time
339 seconds
Digital twins are currently delivering real results, with 70% productivity increases in car manufacturing and 13% improvement in equipment efficiency
Explanation
Cakiroglu provides concrete evidence that digital twin technology is not just theoretical but is delivering measurable results across different industries. He emphasizes that these are proven capabilities being scaled across diverse business operations including home appliances, automotive, and oil refining.
Evidence
Six WEF-recognized lighthouses, 70% productivity increase at car manufacturing plant, 13% overall equipment efficiency increase, increased capacity utilization in oil refining through integrated yield/energy/blending decisions, 95% forecasting accuracy in supply chain management.
Major discussion point
Digital Twins and AI-Powered Manufacturing Systems
Topics
Economic | Infrastructure
Agreed with
– Roland Busch
– Mihir Shukla
Agreed on
Digital twins are delivering real, measurable results in manufacturing
Successful scaling requires strengthening governance, operating systems, talent, and organizational culture, not just deploying more technology
Explanation
Cakiroglu argues that the key to scaling intelligent factory initiatives from pilot projects to enterprise-wide implementation lies in organizational capabilities rather than just technological deployment. He emphasizes the importance of having unified architecture and governance systems to utilize digital capacity across businesses.
Evidence
Platform 360 developed in-house and deployed in more than 40 sites, which standardizes diagnosis and decision-making across businesses and scales execution and decision quality.
Major discussion point
Implementation Strategies and Business Impact
Topics
Economic | Development
Agreed with
– Mihir Shukla
– Roland Busch
Agreed on
Business outcomes should drive technology adoption decisions
Thani Ahmed Al Zeyoudi
Speech speed
171 words per minute
Speech length
1524 words
Speech time
533 seconds
Digital twins and AI are being applied across entire value chains from extraction to distribution, transforming customs clearances and port operations
Explanation
Al Zeyoudi describes how digitalization extends beyond manufacturing to encompass entire value chains including logistics, distribution, and customs operations. He provides examples of how digital transformation has dramatically reduced processing times and improved efficiency in trade operations.
Evidence
UAE’s connection to 250 ports worldwide, customs clearances happening before goods arrive, payments and deposits processed immediately, clearances reduced from days to minutes, AI directing oil and gas production locations, robotics enabling operations without traditional shutdowns.
Major discussion point
Digital Twins and AI-Powered Manufacturing Systems
Topics
Economic | Infrastructure
Successful partnerships require long-term commitment, energy supply, land accessibility, logistics connectivity, guaranteed off-takers, and talent availability
Explanation
Al Zeyoudi outlines the comprehensive ecosystem requirements for successful technology partnerships and industrial development. He emphasizes that all elements must be present simultaneously, as missing any single component can derail entire projects.
Evidence
Examples of projects failing due to missing elements – some countries having land and power but poor logistics making projects unviable. UAE’s experience investing in LDCs within Africa and the systematic approach needed.
Major discussion point
Global Development and Technology Transfer
Topics
Development | Infrastructure
Investment in R&D is critical for customizing solutions to local ecosystems rather than waiting for others to develop technology
Explanation
Al Zeyoudi argues that countries and organizations must invest in their own research and development capabilities to create solutions tailored to their specific conditions and environments. He warns that waiting for others to develop technology results in being left behind in AI and digitalization.
Evidence
UAE’s holistic approach including government appointments, policy development, capacity building, and R&D investments. Two-model approach: deployment/investment model for international technology and consumer market model for piloting projects.
Major discussion point
Global Development and Technology Transfer
Topics
Development | Economic
Disagreed with
– Roland Busch
Disagreed on
Technology transfer and development strategy
Cybersecurity requires government-industry collaboration, employee awareness programs, and continuous development of protection measures
Explanation
Al Zeyoudi emphasizes that effective cybersecurity cannot be achieved in isolation but requires coordinated efforts between government authorities and industry participants. He stresses the importance of employee education and building cybersecurity considerations into development processes from the beginning.
Evidence
UAE’s early start on cybersecurity with proper protocols, employee awareness programs for factory and government workers, and integration of cybersecurity with development processes from the beginning.
Major discussion point
Cybersecurity and Risk Management
Topics
Cybersecurity
Agreed with
– Roland Busch
– Mihir Shukla
– Peggy Johnson
Agreed on
Cybersecurity must be designed from the beginning, not added as an afterthought
Peggy Johnson
Speech speed
184 words per minute
Speech length
936 words
Speech time
305 seconds
Humanoid robots can now operate in unstructured environments built for humans, offering operational efficiency by performing multiple tasks throughout the day
Explanation
Johnson explains that the key breakthrough in humanoid robotics is their ability to work in environments designed for humans rather than requiring specialized setups. This flexibility allows them to perform different jobs throughout the day and navigate human-designed spaces like narrow aisles and varying heights.
Evidence
Contrast with fixed robotics like conveyor belts or single-arm systems. Humanoid robots can go down narrow aisles, reach up high, go down to the floor, and do one job in the morning and something else in the afternoon.
Major discussion point
Humanoid Robotics and Automation Challenges
Topics
Economic | Infrastructure
Humanoid robots address labor shortages in dangerous, repetitive jobs that cause high turnover and injuries, allowing humans to focus on higher-value work
Explanation
Johnson argues that humanoid robots solve critical workforce challenges by taking over jobs that are difficult to fill due to their repetitive, dangerous, or physically demanding nature. She emphasizes that this frees humans to contribute higher-value work while reducing workplace injuries.
Evidence
Jobs that are dull, dirty, and dangerous with high turnover within a year. Aging workforce getting injured more from repetitive manual work. High costs from injuries and difficulty finding workers for manual roles.
Major discussion point
Humanoid Robotics and Automation Challenges
Topics
Economic | Development
Safety remains a critical challenge, with robots currently operating in work cells and requiring functional safety standards to work alongside humans
Explanation
Johnson acknowledges that despite advances, humanoid robots still pose safety risks due to their size and strength, requiring them to operate in controlled environments. She describes the technical requirements for robots to achieve functional safety standards that would allow them to work directly with humans.
Evidence
Robots are 200 pounds and can lift 25 kilograms or 50 pounds. Current requirement to operate inside work cells away from humans. Functional safety requirements including sensing approaching humans and bringing themselves down while maintaining payload.
Major discussion point
Humanoid Robotics and Automation Challenges
Topics
Cybersecurity | Infrastructure
Data protection should be built into robot design, keeping sensitive data on customer premises rather than in external clouds
Explanation
Johnson emphasizes that robots collect vast amounts of sensitive data through their multiple sensors, requiring careful data protection strategies from the design phase. She describes their approach of keeping customer data ownership and processing locally rather than in external cloud systems.
Evidence
Robots have LIDAR, cameras, hearing devices and input much more sensitive data than mobile phones. Company policy of keeping data on customer premises, not owning customer data, and only processing in cloud with customer permission.
Major discussion point
Cybersecurity and Risk Management
Topics
Cybersecurity | Human rights
Agreed with
– Roland Busch
– Mihir Shukla
– Thani Ahmed Al Zeyoudi
Agreed on
Cybersecurity must be designed from the beginning, not added as an afterthought
Jamie Heller
Speech speed
154 words per minute
Speech length
761 words
Speech time
296 seconds
Digital transformation is not optional for companies and can deliver better margins through simple improvements
Explanation
Heller emphasizes that companies face pressure to adopt AI and digital technologies while managing daily business obligations. She suggests that the benefits are so significant that digital transformation has become a necessity rather than a choice for maintaining competitiveness.
Evidence
Reference to companies feeling pressure to ‘get with the program’ while having daily obligations and numbers to make
Major discussion point
Implementation Strategies and Business Impact
Topics
Economic
Companies face a critical decision between starting fresh with new technology versus upgrading existing systems with AI
Explanation
Heller raises the strategic question of whether organizations should implement entirely new technological systems or attempt to integrate AI capabilities into their existing infrastructure. This represents a fundamental choice that affects implementation strategy and resource allocation.
Evidence
Question about whether it’s better to start anew with fresh technology or ‘AI-ify’ what you already have
Major discussion point
Implementation Strategies and Business Impact
Topics
Economic | Infrastructure
Human factors represent the greatest cybersecurity risk in manufacturing systems
Explanation
Heller acknowledges and reinforces the point that despite all technological security measures, humans remain the weakest link in cybersecurity. She emphasizes the importance of well-educated humans in managing these risks effectively.
Evidence
Agreement with Roland Busch’s statement that ‘the most important reason for an attack and a successful one is the human’
Major discussion point
Cybersecurity and Risk Management
Topics
Cybersecurity
Audience
Speech speed
152 words per minute
Speech length
368 words
Speech time
145 seconds
Technology transfer partnerships should focus on building industry rather than just creating markets, particularly in healthcare manufacturing in Africa
Explanation
An audience member from Nigeria’s healthcare initiative emphasizes the need for genuine partnership in industrial development rather than simply selling to African markets. They seek sustainable growth through technology transfer that builds local manufacturing capabilities in healthcare.
Evidence
Nigeria’s presidential initiative for unlocking healthcare value chain through technology transfer, focus on boosting local healthcare manufacturing
Major discussion point
Global Development and Technology Transfer
Topics
Development | Economic
Complex supply chains require strategic timing and holistic approaches for digital twin implementation
Explanation
A textile manufacturing executive managing global supply chains questions when and how to implement digital twins across complex operations spanning from farmers to final customers. They emphasize the need to consider return on capital employed when making these technological investments.
Evidence
Home textile company manufacturing for 60 countries with complex supply chain from farmer to spinning to weaving to processing to cut and sew, deploying 20,000 people with machines that can communicate
Major discussion point
Implementation Strategies and Business Impact
Topics
Economic | Infrastructure
Increased factory automation creates significant cybersecurity vulnerabilities that require proactive addressing
Explanation
An audience member specializing in human-robot collaboration raises concerns about the cybersecurity risks that come with increased automation in factories. They highlight that malware stopping production could cause enormous damage, questioning how companies are addressing these growing risks.
Evidence
Example of malware stopping production causing great damage, reference to working in human-robot collaboration at Italian Institute of Technology
Major discussion point
Cybersecurity and Risk Management
Topics
Cybersecurity | Infrastructure
Agreements
Agreement points
Digital twins are delivering real, measurable results in manufacturing
Speakers
– Roland Busch
– Levent Cakiroglu
– Mihir Shukla
Arguments
Digital twins enable comprehensive simulation and optimization of manufacturing processes before and during production, resulting in 20-25% higher productivity and 20% less energy cost
Digital twins are currently delivering real results, with 70% productivity increases in car manufacturing and 13% improvement in equipment efficiency
Digital twins combined with AI agents create proactive systems that can predict problems and trigger countermeasures across entire supply chains
Summary
All three speakers agree that digital twins are not just theoretical concepts but are currently delivering significant, measurable improvements in manufacturing efficiency, productivity, and cost reduction across various industries.
Topics
Economic | Infrastructure
Cybersecurity must be designed from the beginning, not added as an afterthought
Speakers
– Roland Busch
– Mihir Shukla
– Thani Ahmed Al Zeyoudi
– Peggy Johnson
Arguments
Modern manufacturing systems require end-to-end cybersecurity design from the beginning, as attack angles are now 360 degrees
AI systems should be designed with fixed parameters for critical functions while allowing dynamic reasoning only where appropriate
Cybersecurity requires government-industry collaboration, employee awareness programs, and continuous development of protection measures
Data protection should be built into robot design, keeping sensitive data on customer premises rather than in external clouds
Summary
All speakers strongly agree that cybersecurity cannot be an afterthought but must be integrated into system design from the beginning, with comprehensive approaches covering technical, human, and organizational aspects.
Topics
Cybersecurity | Infrastructure
Business outcomes should drive technology adoption decisions
Speakers
– Mihir Shukla
– Roland Busch
– Levent Cakiroglu
Arguments
Companies should prioritize based on business outcomes (revenue increase, cost reduction, risk reduction) rather than chasing technology promises
Companies should start with data collection and breaking down data silos before implementing comprehensive digital twin systems
Successful scaling requires strengthening governance, operating systems, talent, and organizational culture, not just deploying more technology
Summary
Speakers agree that technology implementation should be driven by clear business objectives and practical considerations rather than pursuing technology for its own sake, emphasizing the importance of foundational capabilities and organizational readiness.
Topics
Economic
Similar viewpoints
Both speakers emphasize that successful industrial and technological development requires comprehensive infrastructure foundations and systematic approaches, with energy and logistics being fundamental prerequisites for advanced manufacturing and technology deployment.
Speakers
– Roland Busch
– Thani Ahmed Al Zeyoudi
Arguments
Economic development follows waves: energy and logistics infrastructure first, then industry, then higher education and healthcare
Successful partnerships require long-term commitment, energy supply, land accessibility, logistics connectivity, guaranteed off-takers, and talent availability
Topics
Development | Infrastructure
Both speakers distinguish between traditional fixed robotics and new AI-powered humanoid robots, emphasizing the superior flexibility and adaptability of humanoid robots that can work in human-designed environments and perform varied tasks.
Speakers
– Peggy Johnson
– Roland Busch
Arguments
Humanoid robots can now operate in unstructured environments built for humans, offering operational efficiency by performing multiple tasks throughout the day
Traditional caged robots perform repetitive tasks precisely but lack the flexibility that AI-powered humanoid robots can provide
Topics
Economic | Infrastructure
Both speakers emphasize the importance of building local capabilities through technology transfer, training programs, and R&D investments rather than simply importing finished solutions, focusing on sustainable development and local empowerment.
Speakers
– Mihir Shukla
– Thani Ahmed Al Zeyoudi
Arguments
Technology transfer should include substantial training programs, with examples of 700 women trained resulting in 525 finding jobs within the first week
Investment in R&D is critical for customizing solutions to local ecosystems rather than waiting for others to develop technology
Topics
Development | Economic
Unexpected consensus
Humans as the greatest cybersecurity risk
Speakers
– Roland Busch
– Jamie Heller
– Thani Ahmed Al Zeyoudi
Arguments
Modern manufacturing systems require end-to-end cybersecurity design from the beginning, as attack angles are now 360 degrees
Human factors represent the greatest cybersecurity risk in manufacturing systems
Cybersecurity requires government-industry collaboration, employee awareness programs, and continuous development of protection measures
Explanation
Despite the panel’s focus on advanced AI and robotics technologies, there was unexpected consensus that humans remain the weakest link in cybersecurity. This is surprising given the high-tech nature of the discussion, yet all speakers acknowledged that human factors, not technological vulnerabilities, pose the greatest security risks.
Topics
Cybersecurity
Technology should not be pursued for its own sake
Speakers
– Mihir Shukla
– Roland Busch
– Levent Cakiroglu
Arguments
Companies should prioritize based on business outcomes (revenue increase, cost reduction, risk reduction) rather than chasing technology promises
Companies should start with data collection and breaking down data silos before implementing comprehensive digital twin systems
Successful scaling requires strengthening governance, operating systems, talent, and organizational culture, not just deploying more technology
Explanation
Unexpectedly, technology leaders and executives consistently warned against technology-driven approaches, instead advocating for business-outcome-driven strategies. This pragmatic consensus from technology advocates suggests a mature understanding of implementation challenges and the importance of organizational readiness over technological sophistication.
Topics
Economic
Overall assessment
Summary
The panel demonstrated remarkably high consensus across key areas: the proven value of digital twins in manufacturing, the critical importance of cybersecurity-by-design, the need for business-outcome-driven technology adoption, and the requirement for comprehensive infrastructure and organizational capabilities to support advanced manufacturing technologies.
Consensus level
Very high consensus with strong alignment on both technical capabilities and implementation strategies. The speakers, despite representing different sectors (technology vendors, manufacturers, government), shared similar perspectives on the maturity of digital twin technology, the practical challenges of implementation, and the fundamental requirements for successful technology deployment. This high level of agreement suggests that the industrial AI and robotics sector has reached a level of maturity where best practices and critical success factors are well understood across different stakeholder groups. The implications are positive for widespread adoption, as the alignment between technology providers, users, and policymakers creates a conducive environment for scaling these technologies globally.
Differences
Different viewpoints
Implementation approach for digital transformation
Speakers
– Roland Busch
– Mihir Shukla
Arguments
Companies should start with data collection and breaking down data silos before implementing comprehensive digital twin systems
Companies should prioritize based on business outcomes (revenue increase, cost reduction, risk reduction) rather than chasing technology promises
Summary
Busch advocates for a technology-first approach starting with data infrastructure and machine monitoring, while Shukla emphasizes a business-outcome-first approach that prioritizes financial returns before selecting appropriate technology solutions
Topics
Economic | Infrastructure
Technology transfer and development strategy
Speakers
– Roland Busch
– Thani Ahmed Al Zeyoudi
Arguments
Economic development follows waves: energy and logistics infrastructure first, then industry, then higher education and healthcare
Investment in R&D is critical for customizing solutions to local ecosystems rather than waiting for others to develop technology
Summary
Busch suggests following a sequential development model where countries must complete foundational infrastructure before advancing to industrial technology, while Al Zeyoudi argues for immediate R&D investment to avoid being left behind in AI and digitalization
Topics
Development | Infrastructure
Unexpected differences
Role of human workforce in automation
Speakers
– Peggy Johnson
– Thani Ahmed Al Zeyoudi
Arguments
Humanoid robots address labor shortages in dangerous, repetitive jobs that cause high turnover and injuries, allowing humans to focus on higher-value work
Digital twins and AI are being applied across entire value chains from extraction to distribution, transforming customs clearances and port operations
Explanation
While both speakers advocate for automation, Johnson focuses on humanoid robots as solutions to labor shortages and worker safety, while Al Zeyoudi discusses replacing non-skilled labor entirely with robotics in construction, representing different philosophies about human-robot collaboration versus replacement
Topics
Economic | Development
Overall assessment
Summary
The panel showed surprising consensus on the importance and current reality of digital transformation in manufacturing, with disagreements primarily centered on implementation strategies and development approaches rather than fundamental concepts
Disagreement level
Low to moderate disagreement level. Most conflicts were methodological rather than philosophical, suggesting that while speakers agree on the destination of digital transformation, they have different preferred paths to get there. This indicates a maturing field where basic concepts are accepted but best practices are still being debated.
Partial agreements
Partial agreements
Similar viewpoints
Both speakers emphasize that successful industrial and technological development requires comprehensive infrastructure foundations and systematic approaches, with energy and logistics being fundamental prerequisites for advanced manufacturing and technology deployment.
Speakers
– Roland Busch
– Thani Ahmed Al Zeyoudi
Arguments
Economic development follows waves: energy and logistics infrastructure first, then industry, then higher education and healthcare
Successful partnerships require long-term commitment, energy supply, land accessibility, logistics connectivity, guaranteed off-takers, and talent availability
Topics
Development | Infrastructure
Both speakers distinguish between traditional fixed robotics and new AI-powered humanoid robots, emphasizing the superior flexibility and adaptability of humanoid robots that can work in human-designed environments and perform varied tasks.
Speakers
– Peggy Johnson
– Roland Busch
Arguments
Humanoid robots can now operate in unstructured environments built for humans, offering operational efficiency by performing multiple tasks throughout the day
Traditional caged robots perform repetitive tasks precisely but lack the flexibility that AI-powered humanoid robots can provide
Topics
Economic | Infrastructure
Both speakers emphasize the importance of building local capabilities through technology transfer, training programs, and R&D investments rather than simply importing finished solutions, focusing on sustainable development and local empowerment.
Speakers
– Mihir Shukla
– Thani Ahmed Al Zeyoudi
Arguments
Technology transfer should include substantial training programs, with examples of 700 women trained resulting in 525 finding jobs within the first week
Investment in R&D is critical for customizing solutions to local ecosystems rather than waiting for others to develop technology
Topics
Development | Economic
Takeaways
Key takeaways
Digital twins combined with AI are delivering measurable results now, with 20-25% productivity gains and 20% energy cost reductions in manufacturing
Humanoid robots are breakthrough technology for unstructured environments, addressing labor shortages in dangerous, repetitive jobs while allowing humans to focus on higher-value work
Success requires prioritizing business outcomes (revenue increase, cost reduction, risk reduction) over technology promises, with some implementations showing $200 million in cost savings
Scaling AI and robotics requires strengthening governance, operating systems, talent, and organizational culture, not just deploying more technology
Economic development follows predictable waves: energy and logistics infrastructure first, then industry, then higher education and healthcare
Cybersecurity must be designed from the beginning as an end-to-end solution, as modern manufacturing systems face 360-degree attack angles
Technology transfer and global partnerships require long-term commitment, energy supply, land accessibility, logistics connectivity, guaranteed off-takers, and talent availability
Companies should start with data collection and breaking down data silos before implementing comprehensive digital twin systems
Resolutions and action items
Companies should begin with data layer collection from manufacturing machines to enable predictive maintenance and reduce downtime
Organizations should tear down data silos and create data warehouses/lakes to enable AI-powered insights across operations
Businesses should target specific high-impact areas (inventory management, order processing, container planning) for quick wins to fund broader AI initiatives
Governments and industry must collaborate on cybersecurity protocols and employee awareness programs
Technology providers should offer free licenses and training programs to develop talent, particularly in developing regions
Humanoid robot developers must achieve functional safety standards to enable operation outside work cells alongside humans
Unresolved issues
How to balance the need for dynamic AI reasoning with fixed parameters for critical safety functions in manufacturing systems
The optimal timing and sequencing for implementing digital twins in complex, multi-stage supply chains spanning multiple countries
How to effectively scale pilot AI projects to enterprise level while maintaining quality and consistency across diverse business units
The challenge of developing local R&D capabilities versus relying on technology transfer partnerships in developing economies
How to address the skills gap and workforce transition as AI and robotics replace traditional manufacturing jobs
The long-term implications of humanoid robots moving from enterprise environments into homes and society
Suggested compromises
Start with brownfield applications (upgrading existing systems) rather than requiring complete greenfield implementations for digital twins
Use hybrid approaches combining on-premise data storage with cloud processing for AI systems to balance security and functionality
Implement both deployment/investment models (using international technology) and consumer market models (becoming regulatory labs for pilot projects) simultaneously
Focus on specific manufacturing areas (like logistics and material handling) first before expanding humanoid robots to more complex manufacturing tasks
Combine technology partnerships with in-house capability building rather than choosing one approach exclusively
Design AI systems with locked-in parameters for critical functions while allowing dynamic reasoning for optimization tasks
Thought provoking comments
I think you almost have a redesigned security first. Exactly. It can be afterthought, it cannot be afterthought, this is, the question is so important.
Speaker
Mihir Shukla
Reason
This comment reframes cybersecurity from a reactive add-on to a fundamental design principle. It challenges the traditional approach of implementing security measures after system development and emphasizes the critical importance of security-first thinking in AI and robotics deployment.
Impact
This comment immediately resonated with other panelists, with the Minister reinforcing the point about starting cybersecurity ‘hand-in-hand with the whole development.’ It shifted the discussion from technical solutions to fundamental design philosophy, elevating cybersecurity from a technical concern to a strategic imperative.
I wouldn’t start with technology at all. Is digital twin the right answers? Maybe not. I would look at, there are only three outcomes you can have. You can either increase revenue, reduce managed cost or reduce risk… chase purpose not promises of technology
Speaker
Mihir Shukla
Reason
This is a profound reframing that challenges the technology-first mindset prevalent in industrial discussions. Coming from a technology company CEO, it’s particularly striking as he advocates for business outcomes over technological solutions, providing a practical framework for decision-making.
Impact
This comment fundamentally shifted the conversation from ‘what technology should we use’ to ‘what business problem are we solving.’ It provided a clear methodology for the textile manufacturer’s complex question and influenced Roland’s subsequent response to focus on practical, data-driven approaches rather than comprehensive digital twin implementations.
There is no amount of human workforce that could do that. And it saved them $200 million in inventory costs… This is a phenomenal time to be doing this.
Speaker
Mihir Shukla
Reason
This comment illustrates the transformative scale of AI capabilities beyond human capacity, not just efficiency improvements. The specific example of managing 685,000 items across 32 warehouses four times daily demonstrates AI’s ability to operate at scales impossible for humans, fundamentally changing what’s possible in operations.
Impact
This concrete example with specific metrics ($200 million savings) shifted the discussion from theoretical benefits to tangible, measurable outcomes. It reinforced the urgency expressed by Jamie’s follow-up that ‘it’s really not optional,’ moving the conversation toward implementation imperatives rather than exploratory possibilities.
The most important reason for an attack and a successful one is the human.
Speaker
Roland Busch
Reason
This succinct observation cuts through all the technical cybersecurity discussions to identify the fundamental vulnerability. It highlights the paradox that in our rush to automate and digitize, human factors remain the weakest link, requiring a holistic approach to security that includes human training and awareness.
Impact
This comment served as a powerful conclusion to the cybersecurity discussion, validating the Minister’s emphasis on employee awareness and training. It brought the technical cybersecurity conversation full circle back to human factors, reinforcing that technology solutions alone are insufficient.
Finance is critical, for sure, but many countries are having the finance, but they don’t have the right solutions or the right diagnosis to sort out the issues… it’s too late to wait for someone else to do it, because you’re going to be out of the game.
Speaker
Thani Ahmed Al Zeyoudi
Reason
This insight challenges the common assumption that financial resources are the primary barrier to technological adoption. It reframes the challenge as one of strategic thinking and problem diagnosis rather than capital availability, while emphasizing the competitive urgency of AI adoption.
Impact
This comment expanded the discussion beyond technical and financial considerations to include strategic thinking and national competitiveness. It influenced the subsequent conversation about R&D investments and the importance of customization to local conditions, adding a geopolitical dimension to the industrial AI discussion.
In this exuberance of AI, sometimes we replace natural intelligence with artificial stupidity and that’s a wrong answer.
Speaker
Mihir Shukla
Reason
This witty but profound observation warns against the indiscriminate application of AI technology. It suggests that the excitement around AI capabilities can lead to poor implementation decisions, emphasizing the need for thoughtful integration that preserves human judgment where appropriate.
Impact
This comment introduced a note of caution into an otherwise enthusiastic discussion about AI capabilities. It reinforced the importance of thoughtful design principles in AI implementation and supported the broader theme of security-first, purpose-driven technology adoption that had emerged throughout the panel.
Overall assessment
These key comments fundamentally shaped the discussion by shifting it from a technology-showcase format to a more nuanced, business-focused conversation about implementation realities. Mihir Shukla’s business-outcome framework and Busch’s human-centric security insight particularly influenced the panel’s direction, moving from ‘what’s possible’ to ‘what’s practical and safe.’ The Minister’s strategic perspective added crucial context about national competitiveness and long-term thinking. Together, these comments created a more balanced discussion that acknowledged both the transformative potential and the implementation challenges of industrial AI, emphasizing the need for thoughtful, security-conscious, business-driven approaches rather than technology-first solutions.
Follow-up questions
How to effectively scale AI and digital twin technologies from pilot studies to enterprise-wide implementation across diverse manufacturing operations
Speaker
Levent Cakiroglu
Explanation
Levent emphasized that scaling from pilot to enterprise level is not just about deploying more technologies, but requires strengthening governance, operating systems, talent, culture and leadership – suggesting this remains a complex challenge requiring further research
When and how to prioritize digital twin implementation in complex, multi-stage supply chains with global operations
Speaker
Dipali Goenka (audience member)
Explanation
This textile manufacturer’s question about timing and approach for digital twin deployment in a complex supply chain from farmers to global customers represents a common challenge that needs more specific guidance
How to design comprehensive end-to-end cybersecurity solutions for interconnected manufacturing systems that can detect and stop attacks in real-time
Speaker
Julia (audience member) and Roland Busch
Explanation
The discussion revealed cybersecurity as a critical challenge requiring security-first design approaches, with Roland noting that attacks are inevitable but can be stopped if detected quickly enough
What specific criteria and ecosystem requirements define successful technology transfer partnerships, particularly in developing markets like Africa
Speaker
Eniton Tadjoshu (audience member) and Thani Ahmed Al Zeyoudi
Explanation
The discussion identified multiple critical factors (long-term stability, energy, land access, logistics, off-takers, talent) but suggested more research is needed on how to effectively address these interconnected challenges
How to determine optimal business cases and ROI calculations for AI and robotics investments across different manufacturing contexts
Speaker
Dipali Goenka (audience member) and Mihir Shukla
Explanation
While examples of significant cost savings were provided, the question of how to systematically evaluate and prioritize AI investments based on business outcomes requires further methodology development
How to balance dynamic AI decision-making capabilities with fixed safety parameters in manufacturing robotics
Speaker
Mihir Shukla
Explanation
Mihir’s point about not replacing ‘natural intelligence with artificial stupidity’ and the need to lock certain parameters while allowing dynamic reasoning in others requires further research on optimal AI system design
What are the most effective approaches for workforce transition and retraining as manufacturing becomes more automated
Speaker
Peggy Johnson and Mihir Shukla
Explanation
While success stories were shared about training programs, the broader challenge of workforce transformation in the age of humanoid robots and AI requires more comprehensive research and solutions
Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.
Related event
World Economic Forum Annual Meeting 2026 at Davos
19 Jan 2026 08:00h - 23 Jan 2026 18:00h
Davos, Switzerland