Hugo Scott-Gall: Today, I am delighted to have with me Claudio Simão, the Chief Technology Officer of Hexagon, a global leader in sensors, software, and autonomous solutions. As well as being the CTO, Claudio is President of Hexagon’s Innovation Hub and has a number of patents in sensor technology. Hi, Claudio.
Claudio Simão: Hello, pleased to be here.
Hugo Scott-Gall: Also with me is Andy Siepker, William Blair’s investment management industrials analyst. Andy, correct me if I’m wrong but you have no patents.
Andrew Siepker: None yet, Hugo; but certainly many patents pending, of course.
Hugo Scott-Gall: Absolutely, of course. Now, we are going to discuss today the fourth industrial revolution and the central role of all things digital in it.
Claudio, could you give us maybe a slightly fuller intro into Hexagon than I gave and then expand into really defining what Industry 4.0, what it means and how it ranges from all the way from augmented reality to robots to cobots, exactly kind of what it is?
Claudio Simão: Yes. So Hexagon, as you mentioned, is a major player. We started something around 20 years ago as a leader in sensors. So that means developing sensors, we had to understand the application and we are very close to data. So data capture has been our, let’s say, focus 20 years ago, and then we evolved it to more software centric solutions developing workforce, addressing workforce for very important vertical applications. And finally, around 10 years ago, we started to dedicate ourselves to enterprise end-to-end solutions.
So let’s say our natural evolvement from sensors capturing data, capturing the real world, streaming the real world, modeling, simulating, detecting anomalies and deviations, and then feeding back corrective actions into the real world from the digital world is a natural, let’s say, path into digital transformation that in a more holistic perspective is encompassing Industry 4.0 basically.
So if you want me to define Industry 4.0, it’s – and if you permit to start with digital transformation, as I said in a more holistic perspective, it’s all about utilization of new digital technologies and emerging digital enablers, added up to core technology and expert systems in order to provide new streamlined solutions to existing processes or creating fully new solutions in a more disruptive manner for enhancing productivity and quality. This would be a definition of Industry 4.0.
In this context, that means Industry 4.0 can be seen as digital transformation for the industry and more specifically for manufacturing. This is how we see Industry 4.0.
Hugo Scott-Gall: And if you put together all the effects of that, is that going to be the missing link where in the productivity puzzle when people say, look, we’re having a technology revolution but it hasn’t really changed productivity, in fact most DM countries have a productivity problem, do you think that the combination of all the different elements of Industry 4.0 will lead to something of a productivity revolution that is it’s maybe just about to come that we should be a little more patient but actually that the combination of things will transform the efficiency, accuracy, productivity within the industrial sector and also outside of it in terms of its broader footprint?
Claudio Simão: Yes, we strongly believe that Industry 4.0 for the factories, for the industry, is a major, let’s say, movement to improve the productivity, quality, safeness, security simply because today, the processes are fragmented. And by the fact that you are interconnected, interoperating fragmented processes, siloed processes and data, will permit a major improvement in productivity, quality, security, and other, let’s say, elements of a factory, of the industry. So we believe strongly that there will be major gains in productivity.
Andrew Siepker: I mean the opportunity that you describe is big and is very powerful for not only companies but for economies in general. But, I guess, where do you see companies when they’re starting on the digital journey? Where do they stumble or what are some of the challenges they face throughout the digital transformation as they undergo that transformation?
Claudio Simão: Yes, it’s clear that there are many pitfalls and challenges for industrial companies similar of what happens at the threshold of prior technology cycles, right, or the Industry 3.0 or 2.0.
So, first, there are organizational, cultural challenges. That means we have to rewire the organization to address the different approaches of Industry 4.0. And there are technical challenges, right?
So if you consider that, for us, digital transformation and Industry 4.0, as I mentioned before, is utilization of new digital technologies to enable different application or streamline the existing application. So, now, we have to use backend integration, enhanced connectivity with a faster, let’s say, wireless or wired networks as in cloud computing, advanced analytics, AI, visualization, augmented reality and all these new enablers that we have or even actuators like robots and drones.
So there are technology challenges, but also there are cultural and organizational challenges because the mindset for the implementation of these enablers, of new technologies are not there, the organizations are not prepared for this.
Andrew Siepker: Yes. That’s helpful. Just as a follow-up to that, I’ve seen a lot of discussion about just the amount of industry data that we’re generating. But the gap between the data that’s being generated and the data that is actually being used efficiently, that gap continues to widen.
So, is that something that companies are just overwhelmed with the amount of data and they don’t know what to do with it? Is that something that you’re seeing? And how do customers – how do you help customers handle that challenge?
Claudio Simão: Yes. Frankly speaking, the data platform is one of the most important enablers for digital transformation and Industry 4.0. I guess in the beginning, we thought that we could have data warehouses, and organizing the warehouses would solve the problems.
And now, we are recognizing that. And then there was a solution for the data lakes, and people understand now that it’s impossible to scale this simply because the data, even structured data, let alone unstructured data, the data is not contextualized, is not stagnant for the applications.
So the data platform hurdle, let’s say, the data platform challenge is major – maybe it’s the most important one besides the cultural and organizational one because as the data is not contextualized and the applications are more and more going to real-time, we simply cannot use historians. What we can use is more time series, but this is very limiting because the historians give us the way to data mine and train systems for improvement, for self adapting in order to deliver really productivity and quality we mentioned before.
So data platform is something that we are investing a lot. And there are a lot of new technologies associated with this and associated with machine learning and data mining, and how we structure, let’s say, the data warehouse as contextualized for different applications.
Hugo Scott-Gall: And, Claudio, this is a much more complex complicated world, which if you have the skills, you can thrive in it. But not everyone is going to win in this world. You need to have the data, you need to understand data, therefore you need to have very good analytics. But what do you see is the defining characteristics of those who are going to succeed in this world and those who aren’t? How would you kind of separate the winners and the losers?
Claudio Simão: So companies have to – they have to prepare. Who has started early will benefit more in this game. And from the organization portion of this, it’s really a different mindset. So companies, they have to put, let’s say, the correct organization to make happen. They have to empower and entrust new leadership to implement a new solution related to digital transformation and Industry 4.0.
Regarding to data, this I mentioned, maybe data, how you treat data is the most challenging one. And I do believe that some companies will dedicate themselves to supply consulting and support to the companies that have not sorted this out.
So I guess the answer is a mix. The companies that will be the early successful companies who is the companies that started early preparing the technology enablers and preparing data platforms but also rewiring the organization to, let’s say, to embrace IoT and the other technology, including AI for example. So it’s – I guess it’s all about preparation for the new, let’s say, era.
Andrew Siepker: So, Claudio, maybe bringing it to Hexagon and how Hexagon can help companies, realize companies’ Industry 4.0 vision, you talked a lot about at the intro now offering really end-to-end solutions. Could you maybe talk about how you see the company’s role in bringing Industry 4.0 to life?
Claudio Simão: Yes, sure. At Hexagon, Industry 4.0 is what we call smart factory. So smart factory is one of digital transformation solutions. We call smart X solution; smart factory, smart mining, smart agriculture, smart construction, and smart (CD) and so on and so forth.
So we have already – for example, for mining, what we call life of mine is from peat to port. We have a complete end-to-end solution. And when I say end-to-end, I mean interoperable solutions. That means you can analyze this for you, you can plan and schedule for multi years from fleet management to processing and so on and so forth.
Same for agriculture. Again, for industries in the scope of Industry 4.0, we call it smart factory. And as I said, this is stepwise. When we start integrating a manufactory cell that means we have a machine tool, a conveyor belt or a robot, and a measuring machine integrated and interoperating. That means if there are deviations in machine, we can detect it, we can measure, and we can correct in real-time. And we can use historians, that means a historical information, to analyze trends, to detect patterns, and with these we can predict the next step improving the process.
So this is one, let’s say, step, one building block of a smart factory. In the moments for example that we integrate CAE/CAD/CAM, that means that what you simulate, you feed back to the design, and you can simulate again and so on and so forth. You can prepare the production, or we can measure the quality, the dimension, deviation in the production, and we can sit back the model in the CAD/CHI. This is already a first step. It’s not a real time but it’s going to be real time.
Andrew Siepker: When you look at some of these solutions in manufacturing feedback, loops, the types of things that you described, simulation, I mean any numbers you can put to that just in terms of sort of someone who’s best-in-class, the type of productivity gains they can see from deploying some of these technologies?
Claudio Simão: Yes, we are going into an end-to-end process, right? In the moment that we have all this smart factory, think about the complete factory, all interoperable, that means all the processes talk to each other and optimize each other. This is a holy grail. So we are not there for sure, and we will go step by step.
But if we consider just one solution that is what we call the notification system, that means if there are deviation, the process, the system feedback this and we can support decisions to this, to, let’s say, change the processing order to optimize the process and quasi real-time, it’s one solution that we are deploying. In the proof of concept, we have benefits going to the level of 20 percent of improvement in the process. So this is a number that I could tell you now. It’s a proof of concept. But the cell we are sending already. But what I’m talking now, the number I’m referring now is it’s out of the proof of concept that we run six months ago or less.
But then we have also solutions in mining, agriculture, in plants that have already shown improvements, double these improvements in different processes.
Hugo Scott-Gall: Just as you are talking, Claudio, I was thinking what does this mean for developed market, so I mean European, U.S. companies versus emerging market companies? Does this kind of level the playing field? Or has it may be raised entry barriers and make it perhaps harder for emerging market companies to climb up the value-added curve that actually this is adding sophistication and complexity to manufacturing processes that you need to have a certain level of skill to stay in the game and that may actually favor developed market companies?
Or maybe that’s the incorrect argument that actually the other side is this will accelerate emerging market companies, other curve, and make them more potent competition for U.S., European, and other DM, Australian, et cetera, companies? Which do you think is right? Is it leapfrogging opportunity for EM? Or is it actually a reinforcement of competitive advantage for developed market companies?
Claudio Simão: Yes. Personally, I believe that emerging markets will leapfrog complete cycles of technology with Industry 4.0 and with digital transformation in general because the domain knowledge is fundamental. It’s not – there is no agnostic approach to this because if you have to train models from the machine learning, artificial intelligence techniques perspective, and if you have to configure systems in a first level and then the system will self-adapt, you need domain knowledge.
But in the moment that you have this, this is all about data, depends how much data you have available to make your systems in a self-adaptive approach to get it better. And if you consider this that we are supplying systems to specific vertical business, for the emerging markets, they will – there will be international companies, right, playing.
And for the emerging market, they will leapfrog the cycle because they will receive the first level of configurations ready. But at the same time it depends on the industry. If the domain knowledge is, let’s say, contained for some few players, this will increase barriers of entry as you said.
Hugo Scott-Gall: Yes. And just sort of thinking some more around this theme. Does the step change in technology which may well enable faster production, more accurate production, that could mean a production sits much closer to where consumption is and therefore this in a sense that these – the confluence of these technologies that could make manufacturing a broader set of industries look very different may well enable – they may well actually deglobalizing. They could be globalizing in a sense that as you’ve just said that more and more companies – all companies around the world will eventually have access to these skills that will transform how they make and maybe even what they make. But actually it could be deglobalizing it. It could localize production because it’s – it will be necessary to sit much closer to production.
Is that too grand a statement that this is – it’s a step change and productivity is a step change, therefore inefficiency. But it may well actually become a deglobalizing force, not a globalizing force, because it means that you can actually put your production much closer where a consumption is and that’s what consumers want. Is that too …
Claudio Simão: Yes. Yes.
Hugo Scott-Gall: … broad and sweeping generalization? Yes.
Claudio Simão: No, I do believe in decentralization of processes. That means production will be more settled. Because if you – when we are talking about Industry 4.0, we are considering all the new enablers, including AI. And AI is a major game changer. As you know, it will revolutionize how we address the entire process and integrate systems.
So systems will get self-adapting. And this is accelerating clearly for me. And when you have systems that are self-adapting and if they are interoperative, that means if we can – if your systems architecture is making possible to interoperate the cells, then we – exactly what you said, you get closer to the consumer, right? And this can be anti-globalization trend.
Andrew Siepker: we talk about what Hexagon is doing in AI? I know you’ve launched the Xalt platform and it’s more than just AI; but maybe some the early efforts in the AI space in terms of bringing these solutions to customers. And then longer-term, how do you see AI impacting industry?
Claudio Simão: Yes. In Hexagon, we have a platform, AI platform, basically a fabric where you can train models. We can deploy AI related solutions. It’s just a platform.
But then we have also – we develop many tools to train different models for different applications. We don’t believe in agnostic and holistic approach to AI. I think we think that AI is very, very associated to domain knowledge in different applications. But in Hexagon, we strongly believe in autonomy and autonomy is supported by AI.
So in Hexagon, it’s all about our vision of connected autonomous ecosystems. And we see clearly the building blocks for our vision coming together.
So the world is moving as you said before. It’s moving from a causal, cause/effect, deterministic world for a more correlational and probabilistic world. So data is foundational. And that we see clearly how relevant is new technology cycle of autonomy, of AI will be – is really relevant. If you compare to the prior ones, the PC cycle, the web cycle, mobile and cloud cycle, for example. So we can really see how impactful this will be, and we are preparing for this. So autonomy revolution is coming.
Andrew Siepker: And when you think about AI and the kind of potential to really accelerate the development there and unleash that potential, is there something that’s missing? Or do we have everything we need to really see AI take off from here? And particularly, I guess I’m thinking about now coming into a 5G world where you have even more connected devices, more data, less latency, more speed, computing power grows even faster with time. So is there anything missing at this point for AI to really take off?
Claudio Simão: I think for AI, it’s more about data. This is why we are investing so much in our data platform, how we organize the data, how we do the data curation, the data mining, and how we are going to use data historians in real time.
When you think real time, 5G is really – besides other enablers, 5G will make possible many real-time low latency, ultra reliable, massive machine-to-machine communication use cases possible, right? So 5G is a clear accelerator for Industry 4.0 because it will enable many use cases that we simply cannot do now. And this goes from real-time process using historians data as I said because you want to do the data mining in real-time. And the second one – an edge computing for sure, that means machine-to-machine communication.
And second – and going to, let’s say, augmented reality, latency is a very critical for augmented reality applications or virtual reality application as you know. And so, it’s a – 5G is a major, let’s say, enabler.
But if we come back to AI, you asked what is missing for AI. So as I said, first is data. And when I say data is not simple data sitting there in a data lake without any tagging, without any correlation. So this is very limited. We need a new – we need really a paradigm shift from how we treat data, how we architect data. And this is exactly what we are working a lot at Hexagon and I see others. I have a good network in the technology community. So I see others also discussing this a lot. And I guess this is the most important hurdle, meaningful data, actionable data, not only structured data sitting there in a data lake. So I would say this is the most important thing.
And then it comes to the other elements that I say, it’s cultural, it’s organizational people have to embrace. People don’t – things will not happen from day to night overnight because technology takes time. I always say this here to my group TTT, technology takes time. You have to invest.
But in Hexagon, we are already using AI to develop products, not mandatory embedded in the product but to develop products.
We have many process optimization for industrial application in our solutions and mainly metrology already using AI. We have automatic dispatching for police, fire, and ambulance in our public safety solution. We have an optimization of fleet management in mining, construction, agriculture already deployed using AI.
For example, image classification in agriculture to detect where you have to fertilize, where we have to deploy pesticides is already using AI. And I have a lot of use case; predictive maintenance of machinery, mobile sets, and there are a lot of things that we are doing with AI in Hexagon.
Hugo Scott-Gall: I guess a big risk here is cybersecurity. But if everything is digitalized and I’m relying on fast, efficient communication technologies, hacking cyber risk must be forefront. How do you think about that? And how much you worry about that?
Claudio Simão: Yes, we worry a lot, frankly speaking. So as we say and this is one of the mantras of our Xalt platform. So Xalt is all about merging, fusing OT with IT. So, so far, all the cybersecurity, let’s say, not all, but most of the cyber security worries and development was associated with IT. But in the moment that we fuse, we bring together IT with OT, then we have another entry point for invasions, let’s say. And we should also develop our products that our cybersecurity by design. And this is what we have now in our best practices for product development most hardware and software. This is one element.
The other element is enabling edge computing. And when I say edge computing, it’s not the only edge, but it’s cloud for orchestration and edge-to-edge for choreography again. So we don’t believe that is everything is cloud. We have a lot of edge computing, a lot of distributed computing architectures coming in Industry 4.0, in digital transformation solutions.
So if you have a lot of edge computing, you have another, let’s say, weakness of your solution of your architecture because the risk can come from the edge. If you collect in an interoperable way, a lot of data from the edge, you can have another, let say, point of inclusion.
And our solution for this is developing, let’s say, edge authentication. That is a technique that permits you lighting a distributed ledger, permits you to protect the edge. And to make sense when you orchestrate to the cloud, you have to authenticate the edge to your cloud. So this is a kind of a protection with the distributed authentication.
Hugo Scott-Gall: Well, Claudio, I want to do is say a big thank you from me for taking the time and talking about so many different things and so many fascinating things. So it’s a big thank you for me.
Andrew Siepker: Claudio, yes, also thank you …
Claudio Simão: No, it’s – as I said in the beginning, it’s a pleasure to be here.
