What does reshoring mean to you? Depending on where you’re standing, whether its Asia, North America, or in developing countries in Western Europe, reshoring means different things to different people. On this episode of Manufacturing Matters, Sven Dharmani, global advanced manufacturing and mobility supply chain leader at Ernst & Young (EY) joins TECH B2BMarketing’s Winn Hardin and Jimmy Carroll to talk about how complex modeling is helping companies develop both long-and short-tail supply chains, and how integrators and technology suppliers must take advantage. He also talks about the potential for industrial automation across various sectors and the state of international manufacturing markets.

Jimmy Carroll: [00:00:07] Hi everybody. My name is Jimmy Carroll. I’m the vice president of operations at Tech B2B Marketing. We’re here at the Manufacturing Matters podcast, and we have the pleasure of being joined by Sven Dharmani of Ernst and Young, now known as EY. Sven, thank you so much for taking the time. Really appreciate it.

Sven Dharmani: [00:00:24] Thank you for having me here.

Jimmy Carroll: [00:00:26] Of course. Yeah. So for those maybe outside of the professional services, accounting firm space, could you tell us a little bit about what EY does and what you do there?

Sven Dharmani: [00:00:36] Yeah. EY is a multifunctional firm. We are most famous for our audit and tax practice, but we also do transactions and consulting, and consulting is now almost 35%, 40% of the business. So it’s growing really rapidly over the last decade. I joined EY. I’ve been in consulting 25 years. I joined EY by about 14 years ago when we were just restarting the consulting business. And it’s grown tremendously. I think it’s over like 16 billion now. So a very large consulting firm. But our legacy and our heritage is audit and tax. So that’s what we are most famous for. And I think that was part of the rebranding effort a few years ago. Was to make people aware that we’re a multidisciplinary firm. For myself, I started my career in the industry in automotive and then moved into consulting. Twenty-eight years experience altogether, primarily in supply chain. I run our America supply chain for our advanced manufacturing and automotive and industrial sectors. So that’s my responsibility. We have a pretty large team in supply chain and manufacturing that focuses on that. And this is where I’ve primarily spent all my career in discrete manufacturing and then also in process manufacturing. So I can cut across both the types of manufacturing processes, a lot of experience with automotive, industrial clients, chemicals, and so on and so forth. So I’ve had the privilege of working and doing projects in six continents barring Antarctica.

Winn Hardin: [00:02:22] Speaking to that point, I believe right before the show started, we were just talking about a recent trip you just came back from over in Asia. And of course everyone in North American manufacturing and industrial markets is laser focused, always has been on happenings that are happening in the Asian continent, especially individual players and how they interact. So can you share with us a little bit about those discussions?

Sven Dharmani: [00:02:42] Yeah, absolutely. And Asia continues to be a big growth opportunity for a lot of our multinational clients. As we were talking, China is huge in terms of its market potential. The middle class has gone from 30 to 300 million over the last decade. So it’s become a really big potential market. So obviously not something that can be overlooked. However, the global trade and global logistics have become much more challenging than they were, say, 10 years ago. If you go back to 2014, global trade was frictionless. You could make products anywhere and move them anywhere. Global logistics were fairly smooth. But in the last decades we’ve seen obviously not just COVID but also things like the Suez Canal not being usable, easily navigable or even the Panama Canal not having sufficient water. It’s causing congestions. And then through the pandemic, we saw the ports globally getting clogged and the logistics time, transfer times getting really extended. So this made supply chains very brittle and lengthy. And a lot of the global clients are looking at, well, how do I compress my supply chain, the length of my supply chain? Because the longer the supply chain, the less responsive I can be to any of the market trends to my customers.

Sven Dharmani: [00:04:07] And for that purpose, they are looking at maybe a bit more of, let’s focus on China, because the market’s big, we need to look at the rest of APAC as well. I need to look at Europe, and I need to look at Americas. And what they’re looking at is how do I compress the length of my supply chain. And if I can take out an ocean leg, which has become very unreliable, and an ocean leg itself is at least 30 days. So if I can take one or two legs out, then I make my supply chain much shorter and I can react to market changes much faster. And I can use technology like AI and machine learning to predict demand better. And I can use natural language processing to sense what’s happening on social media, whether I’m making products that go to consumers or go B to B and then go to consumers. I can use all of that capabilities to model what I need to do. But if my supply chain is still long, it can be agile. So to drive agility they’re fundamentally rethinking what the network looks like. Where are my suppliers? Where do I make? Where do I store? How do I transport? What do I transport? Is it raw materials? Is it finished goods? It’s a work in process, so that I can be much more responsive to my customers.

Sven Dharmani: [00:05:27] And for that reason digital twins and control towers have become very popular because I can model my entire network and I can do a simulation and say, okay, if I’m making a tire and the demand has suddenly grown for SUV tires or large-diameter tires. I was traditionally sourcing it from Factory A and Factory B, but the demand is far beyond the capacity that’s available right now. So what’s my best way to get it to market, and how can I still get it to market in a relatively short amount of time so I don’t lose my market to my competitors? So the simulation, looking at say digital network modeling, which lets you look at the sources and transportation time, those are really valuable. And they help the companies think through: What’s the best way to respond to market changes and to satisfy needs of my customers? How can I keep my order fill rates high? So we are seeing a trend globally across companies to look at, relook at their supply and entire supply network all the way from suppliers to their customers and then think about refreshing it so that they can manage cost but also sort of look at sustainability at the same time, which 10 years ago really was more of a reporting field.

Sven Dharmani: [00:06:56] It really wasn’t a driver to say, I need to rethink about my manufacturing in my network. But now they’re starting to look at that. And they are looking at the global markets where the growth is and because the nature of the markets are a little bit different. So if you think of Western Europe and North America, they are much more mature markets. They’re stable. But Asia is growing really fast and will continue to grow, even though there might be a short-term slowdown, in the long run between India and China, those are the two biggest consumer markets. And they’re going to drive a lot of growth. So the growth part of the business has to be handled a little bit differently from the mature part of the business, which is steady. So you need to be much more cost competitive in that segment.

Winn Hardin: [00:07:45] That’s interesting because generally the response to a long-tail supply chain that everyone talks about is simply reshoring. It’s just large umbrella. But when you put the complex modeling into effect and you bring in adaptability and flexibility of the whole supply chain, it seems like it’s a much more complex equation. And then if you throw sustainability, energy, power consumption, other aspects into it, it’s not just about product availability but just a much more complex equation to figure out what’s the best fit for one supply chain. Can you talk a little bit more about how companies are adjusting? Manufacturing is not traditionally an agile market, but this seems to imply, things you’re talking about, is a massive need for flexibility and adaptation, which must be very new to many of the industries that you’re dealing with and we work in.

Sven Dharmani: [00:08:39] Yeah, it is. And the pressure has been growing, I’ll say, over the last five years. Even before the pandemic, the pressure started to grow. But if you think about, let’s take an industrial business, there are parts of the business that are very stable. The product demand is stable. It may not be fast growth or high margin, but it’s very stable. And then there are parts of the business that are more dynamic, maybe more profitable, because there’s a short-term opportunity there. So thinking about segmenting the business and saying which parts of my business are relatively stable and which parts of the business are higher growth and more dynamic?  Also how do I look at my margin? So I need to handle my low-margin business differently from my high-margin business. So in case of a low-margin business and if the business is steady. So think of a two-by-two segmentation and you say low-margin business, very, very steady. I can still have a long supply chain because I can forecast my demand. I may have to build more inventory buffer along the way because the logistics is not as reliable, but it’s still very, very stable from a demand perspective.

Sven Dharmani: [00:09:52] So I tackle that segment, which is very stable. I can keep the long supply chain. That’s okay. If I go to the part of the business where the demand is more volatile and I have a high margin, I don’t want to lose market share in that, because that’s where most of my money is coming from. So I may need to really compress my supply chain. And that part of the business, I may need to reshore because I have to produce it very, very quickly. And then for the other parts of the business, I may want to look at how do I process my products differently, so I may keep more work in process. I may do packaging and finishing close to the demand. So for example, I’ll take two examples. I’ll take chemicals as one example. If I make chemicals and I’m packaging them, whether I’m packaging it in a drum or a cartridge or a bulk container, I can keep the materials in large tanks and I can blend it and finish it, like, for example, silicones with the right color additives or the right additives, and then package it into the right packaging quantity, because that’s where it becomes nontransferable.

Sven Dharmani: [00:11:10] When it’s in bulk, you can still sell it to multiple customers. Once it gets into a specific packaging, it’s much harder. So you can do postponement and you can keep more of WIP close to the customers. So then the segments which are low margin but still volatile, that’s the one you really need to think about is, okay, how much do I want to focus on that? Do I want to look at a different business model? Do I want to look at a distributor model, where they take some of the load off me? Yes, I’m going to give away some of the margin, but I’m taking away like 50% of the work of my people to manage the business that only accounts for maybe 5% or 10% of the margin. So you can give away that margin to a distributor, for example. So you really need to strategize on your business and segment your business and think about each. There isn’t one size that fits all.

Winn Hardin: [00:12:08] Agreed. And reshoring even seems like it may not be an appropriate term anymore. And we typically think in very provincial terms: reshoring is to bring it back to whatever domestic market we’re in, whether it’s Europe, North America, Africa, wherever it might be. But we really seem to be talking about just a larger, more distributed manufacturing and supply chain network that is looking at localized demands, especially if you’re a multinational. So do you still use reshoring on a regular basis or have you kind of moved away from that?

Sven Dharmani: [00:12:36] I think reshoring is probably, like you said, it was a term that was maybe coming in a few years ago. We look at reevaluating a network because also what’s changing is it’s no longer a linear supply chain. It’s starting to become more of a network. So think about Sony and LG, or Sony and Samsung. And like 10 years ago or 20 years ago, they were fierce competitors. But now Samsung is also a supplier to Sony. Similarly, the companies that were my competitor are now my customer as well. So because there is so much technology embedded in all the products, it’s evolved. Companies are evolving from a linear supply chain to more of a network. So you really need to reevaluate the network as such, because you can use contract manufacturing. You can use other sources who have flexibility in manufacturing. So if I don’t have enough capacity to meet my surge demand, I can use stalling or I can use contract manufacturing. So that’s why clients are starting to reevaluate the entire supply network. So that’s probably the change that we are seeing.

Jimmy Carroll: [00:14:08] Sven, I wanted to ask about the role of industrial automation technologies in all of this. And if you use the example of Asia, the International Federation of Robotics says that nearly 73% of all robots installed last year were in Asia. And China’s by far the largest user of robots in the world, and I’m guessing that that number will continue to grow based on what we’ve talked about. But in addition to that, you talk about not only production processes but but packaging. And on that same note, market intelligence company Interact Analysis says that the Chinese warehouse automation market will be set to grow at a compound annual growth rate of 13.5% over the next five or so years. So where do you see industrial automation technologies, not just robots but AI, you mentioned simulation, digital twins, machine vision. How big of a role do you see industrial automation having in Asia and beyond?

Sven Dharmani: [00:15:11] I’ll say I see exponential growth in Asia, and I’ll say developed, grown, mature markets, North America and Europe as well, because one thing is consistent globally that we are seeing, which is the war for talent. There isn’t enough skilled labor to support all the needs globally. Even if you think about trucking, the amount of CDL licenses that are being issued have dropped from 10 years ago. The amount of workers available to run the warehouses. So the amount of skilled labor to run more and more complex manufacturing equipment. So because the demand is outstripping the availability of skilled labor, the opportunity for industrial automation is very, very significant. And it’s very helpful, because if you think about it, from a quality perspective and consistency perspective, if you can implement automation, it not only reduces the workload on the people, but it also makes it much more consistent and stable. And we are seeing use of robots and cobots alongside people to help ease the workload.

Sven Dharmani: [00:16:34] And we’re also seeing use of industrial automation, like, think of machine vision for quality control. I mean, you can apply AI and machine learning to machine vision, and they will detect where a product meets or doesn’t meet the quality specifications very, very quickly. You can also use the machine vision to capture a digital twin of the digital replica of the product in itself. So I think Industry 4.0 in general really started 10 years ago. But given the compute power and the bandwidth and also the cloud capabilities, which really got accelerated through COVID, they’re really coming to life now. And I continue to believe that industrial automation is going to help overcome the shortage issue, labor skill shortage, labor shortage issue in Asia. And also the cost of labor issue in North America and Western Europe. So they’re addressing two different problems, but they’re addressing a problem each geography has.

Jimmy Carroll: [00:17:45] Yeah, totally. I love that you say that. Because I think about post-COVID or even during the beginning of COVID, you saw this huge spike in sales figures and automation and machine vision that really just highlighted the importance of automation. But really, the technology was ripe for growth anyway as different types of businesses and industries across the world seek ways to become more efficient and effective and keep pace in this ever-competitive world. Regardless of COVID or any political supply chain type issues, I think automation will only continue to get more important. And even now, if you look at some of the figures of maybe slight contractions in sales last year, that’s only saying the sales figures are down from COVID years. They’re still larger than ever, and they’ll continue to grow and become more important.

Sven Dharmani: [00:18:39] Yeah. And what we are seeing across a lot of our clients is 2023 was the first full normal year, because 2019 was the last normal year; 2020 we had COVID start kicking in in March, February/March. So it was a disruption; 2021 obviously vaccine rollout; people were still on lockdown; ’22 we had the Delta variant, and I’ll say much more of a normalized version of business started in 2023. So we’re starting to see ’23 as almost the first normal year of operations. And, in fact, we are using machine learning and AI for demand forecasting. And in a lot of our cases, we’re telling the clients, let’s separate out ’20, ’21, ’22 as abnormal years and look at all the way through ’19 and then look at ’23 to really predict what your demand is going to look like. Because also not only was the demand volatile through through ’22, but also the supply was so constrained. So you couldn’t even meet the demand. So if you start to look at sales and shipments that you had between ’20 and ’22, they’re going to be all distorted. So we can apply a lot of intelligence to external factors.

Sven Dharmani: [00:20:02] We’re starting to look at economy growth. We’re starting to look at trends. For example, you’re tracking Dow Jones and Nasdaq all-time highs. You’re starting to see interest rates posturing. What will that do for demand for the consumers? And in turn, what will that do for the demand for industries that make products that feed the consumers? So it’s pretty exciting to see where the market’s going to go. And you’re right, Jimmy. I mean, Industry 4.0 automation-type technologies are ripe for consumption, and they’re really starting to grow exponentially in use. And especially with 5G, now you can have connectivity in the plant without actually tapping into the network. So you don’t have to worry about cybersecurity. You can have controls and systems which can be independent of data communication. So I can see what my machines are running at, and I can transmit select pieces of data without infiltrating my network. So I don’t have to worry about cyberattacks and cybersecurity. So definitely a lot more use we can get out of these capabilities.

Winn Hardin: [00:21:19] That’s cool. I actually haven’t heard of a lot of secondary redundant networks solely using 5G but just sharing nonproprietary, unidirectional information probably. So there’s no way to come back into the network. But I think that’s brilliant. I mean, always latency has been a challenge to overcome when it comes to wireless inside-the-plant communications. But to circle back to something you were just talking about, one thing we didn’t even talk about was the normalizing of inflationary pressures as being kind of a proxy. And when we talk about the second half of this year, a lot of economists and people who follow manufacturing are expecting a softening or flattening in the second half of this year. While I’m wondering, well, we see the Federal Reserve, other central banks, if they start to do some easing that always is going to push cap ex. At the same time, we’re seeing inflationary indications kind of really normalize across so many product segments, which would seem to then free up or reduce the volatility fear index for the general consumer. It’s hard to imagine those two massive trends coming together and a flattening in a market that is just beginning a maturity curve where we’re talking about overall industrial automation, AI, and different advanced technologies. Do you have any personal thoughts about the second half of the year or into ’25?

Sven Dharmani: [00:22:33] I think — this is just my personal thoughts — I think we’ll continue to see growth in the use of automation and industrial technologies because it’s very hard to predict where the economies are going to soften and not soften. I’ll say too, if you go back to the 2007 time frame, the global economies sort of rose and fell together. They don’t do that anymore. So what you have is a little bit of a portfolio effect. So if I’m a global multinational company, yes, I may have softening in one part in China as an example. But I mean, the US consumer confidence has been robust, right. So there is a little bit of natural offsetting because the global economies are not moving all together. So at the same time, the pressure for talent is continuing to grow. Most of our industrial hamlet have 5% to 10% unfilled roles. How am I going to run my business if I cannot fill 5% to 15% of my roles? Well, I can use AI. I can use autonomous supply chains. I can use more automation in my warehouse. I can use better technologies.

Sven Dharmani: [00:23:57] For example, if I start using RFID in all my products and have RFID gates, I can receive products autonomously. I don’t need somebody to scan the barcodes. The equipment can go. I can have RFID sensors throughout my warehouse. I can know exactly the location. By the way, I can use that technology for doing my inventory counts. So now I have reduced the amount of people I need to run my business, which helps offset the pressure of, I cannot fill 5% to 15% of my roles. When we looked at these technologies 10 years ago, people were concerned about: Am I going to have to fire a lot of people? Am I going to have to lay off a lot of people? Well, that’s no longer the problem. The problem now is I cannot find enough people to run my business. So this is where technology becomes a true enabler and solves a business issue of how do I continue to grow? How do I continue to improve my margins without negatively affecting my people? So for that very reason, I think it’s going to continue to grow. Industrial automation is going to continue to grow exponentially.

Winn Hardin: [00:25:12] And it seems like you’re addressing it more at the top level, especially on a global national scale, while realizing that in a geographic or local area that could have differences. For example, across all industrial automation, obviously the adoption of technology, the overall market potential for industrial automation is growing. But then we can look in the food packaging and processing industry, they did massive investments when they traditionally have not, during the COVID years. And I think now they’re kind of dealing with trying to take a breath from what we’re hearing from our internal people, for example. And it used to be the semiconductors were cyclical or automotive was cyclical.  I’m not sure if we’re going to return to those kind of normal cycles anymore or if in the case of semiconductors, demand is just diversified across so many units. I mean, there’s like TSMC semiconductors and then there’s rest of the world semiconductors. Again, it seems like you have to really look at the granular level of your market, your industries, and where those end user markets are for an individual company to be able to project and plan, which just screams complexity to me.

Sven Dharmani: [00:26:20] It is. Definitely the complexity of running supply chains and manufacturing is much, much more than what it used to be 10 years ago. You could really take much more of a uniform approach across a product portfolio. Now it’s just like you said, the markets are different. I mean, the trade, tariff, and duties change on a weekly basis, which means my sourcing, my total landed cost for a particular product can change month to month. And also like we talked about, the global logistics is becoming much more challenging between the port congestions that we saw and the Suez Canal, Panama Canal issues. So how do I run my business better? How can I make my manufacturing and supply chain much more predictable? I mean, you saw news, some of the automotive OEMs are looking at insourcing finished vehicle logistics. They’ve never done that for three decades. But because they saw so much volatility and they’re trying to get their cars, their finished product to the dealer, if you can’t get it, they won’t have a sale. So they’re trying to do that and they can’t because the logistics, finished vehicle logistics is suffering and is constrained. So now they’re looking at insourcing it, so that they have a better control of how do I get my finished product, which actually the dealer has a customer order for, and the customer is waiting for the car. How do I get it to the customer in a timely manner? So it’s changing some of the characteristics of how the business ran from 10 years ago to now.

Winn Hardin: [00:28:06] During our whole previous conversation we were talking about the growing network of suppliers. And then while you were talking about that and sharing with us, that was making me think: I wonder if the industrial automation designer teams are up to the task, right? Because if you’re a multinational, you’re a large company, you could develop one automation system and you could scale it. And that was one of the benefits. Wherever you put it, you were going to have brand protection. You were going to have consistency of defect, production yields, things of that nature. But if you’re having to source two or three or four key suppliers in different regions of the world and then maybe have one different type of supply chain for Asia, another one for Eastern or Western Europe, that seems to indicate that the engineers, and this may be too granular for you, Sven, but it just makes me wonder: Are the solutions, which are always custom to some extent, are they flexible enough so that they can adapt? We can have an AMR solution in warehouses that’s going to work just as well in Indonesia as it’s going to in Eastern Europe, for example. Or is there just going to be an exploding demand for engineers to be able to develop one-offs and custom solutions for all these because every operational environment is so different. And then you talk about insourcing and that’s consolidation of suppliers.

Sven Dharmani: [00:29:24] I think you bring up a really good point. For example, if you think about the telecommunication network, in different continents it’s different. So how do I have automation equipment in Europe and North America and China and rest of Asia that is consistent? And so that’s a little bit of the challenge, because if I’m a global corporation, a multinational company, Fortune 100, I want to be able to look at all my plants. I want to be able to look at all my warehouses and say, what’s my productivity, what’s my efficiency? What’s my best plant in the world? I’m going to take my plant manager from the best plant in the world and I’m going to put them in the worst plant that I have so they can help improve. Similarly for the warehouse, this is why there are so many expats moving around is they’re bringing in leading practices from within the company as well. So if I don’t have consistent processes and consistent tools, that becomes really hard to do. And it becomes really hard to compare plants and operations across regions. So I think the approach that the local companies are going to take, for example, whether it’s in China or whether it’s in India, they are going to take maybe a little bit different approach than large global Fortune 100, 200 companies because they need to have visibility across the plants.

Sven Dharmani: [00:30:54] They need to have visibility across the entire supply network. If I’m getting automotive components, I have a product that I’m selling in North America and Europe and Asia, and I’m selling the same product, a luxury SUV, as an example, I may have three different sources for each region or I may have one source for two regions. I want to be able to look across those, and I want to be able to see, okay, is my takt time for my manufacturing for that same SUV with the same similar configuration, how does it compare in a North America plant versus a Europe plant? And then how do I improve it? Because the cost pressures are never going away, and productivity pressures are not going away. So I think for global companies, they have to approach it slightly differently than just a regional local company. They can definitely go with a local provider, a local tech provider, a local telecom and 5G provider, and an integration system that is fit for a purpose. So I think it kind of goes back to what’s fit for a purpose.

Jimmy Carroll: [00:32:07] Sven, years back, around the trade shows, like the Automate trade show and similar events, there was this growing, and in the national media I should say, not just in the industrial space, but there was this growing fear that robots were going to take people’s jobs. And it just really hasn’t transpired that way. In fact, it’s kind of the opposite. Do you see industrial automation as something that can help attract talent, especially young talent in places that maybe they otherwise wouldn’t want to work? I mean, even examples like using a robotic exoskeleton in a warehouse, not just the idea of working next to a cobot, which is cool. Do you see this as something that could attract young talent?

Sven Dharmani: [00:32:54] Absolutely. And the reason also is that what the younger generation wants to do is different. This is one of the reasons why you’re having trouble, especially in certain markets, finding people to work in plants and warehouses. And they are also natively much more conversant with technology than, let’s say, Gen X or boomers. So it is going to attract. So, for example, if you think about I have to maintain a tool, and I have to maintain a die. That happens once a year. Nobody picks up an instruction booklet or manual and says, okay, how am I going to take this die apart? Now imagine if you had augmented reality and you had visual projection, like showing you step by step, how do I take the die apart. Because I have to maintain the die once a year. How do I take the die apart? How do I do it safely without (a) hurting myself or (b) damaging the tool? That’s much less stressful and also a kind of a fun experience because you’re using augmented reality. You’re using Google Glasses or whatever glasses system that you decide to use. But that’s much more native to what I’ll say millennials or Gen Z are comfortable with. So they’re going to be much more easily suited for those kind of roles versus “I’m going to read a book and see how I’m going to take the tool apart.”

Winn Hardin: [00:34:33] Industries like aerospace have been doing this for many years when they’re putting together a turbine, right, highly complex pieces of equipment using augmented reality solutions for that. The trick there, of course, is the amount of programming that went into it meant that only GE, Boeing, and others could afford that. One wonders if we’ll be able to go from CAD file to instructional AR system soon, because they’ve already got CAD modifications to standards that include more, I think SPIF if I’m not mistaken is the right acronym, where the manufacturing tolerances are built in and you can basically just feed in that CAD file into many robotic programming solutions or machine vision, AI-enabled solutions, and it can basically generate the vast majority of the inspection program or the assembly program with those tolerances. It’d be cool if we could take that into the maintenance field. I’m sure that there are people already chasing that with great vigor.

Sven Dharmani: [00:35:27] A couple of other things that come to mind. One is, it’s going to be a transition because the products that have already been prepared, the design was done earlier. Now the products that are being built, you can have those capabilities native into the product development process. So going forward, things being built going forward, industrial products and equipment, can have those capabilities, but things that have been built and industrial equipment has a long life, 10, 20 years, 30 years in some cases useful life, so for them you kind of have to work on how do you retroactively define. But as you’re doing the upgrades, you can kind of build that. The other point you bring up is a very interesting one. Close to my heart is predictive maintenance or maintenance. And we work with many clients where they were doing more of a schedule-based maintenance and not predictive. Preventive maintenance. So you have a client, where they have a weekly process of literally pulling out a power module and looking at the power module because 20 years ago, they still had control power tubes in them. And they would look at the power tube is starting to become darker. So that process has continued. Now it’s all solid state. So what do you look at when you pull the solid state board out? Nothing, right? but the process has continued. And in fact, the solid state devices are digital: zero/one. Either it’s going to work or it’s not going to work.

Sven Dharmani: [00:37:03] And when it starts to have a failure, you can get signals. So now if that power unit has a temperature sensor, a voltage sensor, and a current draw sensor, now I can actually monitor the conditions and say, I know my unit is going to fail because the ambient housing temperature went up, my current draw has increased, and the voltage that it’s operating has dropped because it’s drawing more power. So you can use sensor and IoT capabilities to really go from preventive to predictive, right? Where I can tell. So I don’t have to send a tech out to go look at it every week. I can monitor the parameters. I can have dashboards that start giving me warning that the voltage has dropped below a threshold or the current draw has increased above the threshold. And I’ll give you an example also that you and I, all of us, use on a daily basis in our homes. Ten years ago, when we did the annual maintenance of HVAC, the guy would come and put on a valve and open the system to check the pressure with a pressure gauge. And then close it and then disconnect it. And in that process you’d leak an ounce of freon or refrigerant. You do that 20 times or 10 times, you’ve lost 10 ounces of refrigerant. So yes, your AC coil might freeze over.

Winn Hardin: [00:38:37] Solely because of the maintenance operation.

Sven Dharmani: [00:38:38] Solely because of maintenance. But now you can look at the current draw of the compressor. You can look at the temperature before and after. So you can look at the temperature drop and you can measure other readings and say, okay, my air-conditioning unit is working within the tolerance, within the operating parameters. So it doesn’t have to be intrusive. So you’re not causing a failure in the process of maintenance. Which is, I’m going to leak a little bit of one ounce of refrigerant every time I check it. So I think the enabling technologies, IoT, sensor data, things like that are going to make our operations much more efficient and much more productive going forward. And I think because these are also technologies that interest the younger generation, they’re going to be much more interested in engaging in those versus going out in the field and checking the pressure in the compressor outside in the 100-degree heat.

Winn Hardin: [00:39:49] While creating a massive market for autonomous robotics like Boston Dynamics, Spot, and others that can use nondestructive scanning system to evaluate those older systems that don’t have all of the current voltage monitoring, internal thermocouples, or . . . 

Sven Dharmani: [00:40:07] Another example that comes to mind is a lot of plants have analogue pressure gauges. So I can’t go around and replace every pressure gauge. But what I can do is I can install a high-precision camera, which can translate the needle position into a digital signal and say, okay, the needle is at three. The needle is at four, needle is at five. You can you can teach it through ML to learn the position of the needle and convert that analog signal into a digital signal. So it’s a nonintrusive way, until you go through your equipment replacement process, to digitize your manufacturing in a supply chain. There are lots of other examples. I just wanted to give a very simple example. And cameras are so cheap, 50 bucks for a camera, 100 megapixel camera is like 50 to 100 bucks on an industrial use basis. And with the connectivity, the 5G, the bandwidth that you have now, you can easily get the data. Getting the data in and out used to be a challenge. Now there are so many avenues to get the data safely.

Winn Hardin: [00:41:18] Yeah. And it’s sipping power. So even power supply is probably not a big problem within a plant. Very cheap. In the end, I could see that solution being cheaper than actually replacing the valve. You put labor and cost and component cost into it.

Sven Dharmani: [00:41:30] And you can do that when you go through your equipment overhaul, which is a natural life cycle. But the transition can be eased, I’ll put it that way.

Jimmy Carroll: [00:41:42] Sven, I want to be mindful of your time. But there is one question I wanted to ask before we close up here. Within the context of industrial automation and manufacturing today, have you noticed any big takeaways from recent EY surveys that would be worth mentioning?

Sven Dharmani: [00:41:58] Yeah, it is interesting that you asked that question. One thing that’s very consistent that we’ve seen is a lot of clients get into this pilot purgatory, where they do a pilot, they test a capability, but they get stuck. And the reason we believe that’s happening is they haven’t really clearly defined the problem statement. They haven’t defined the value proposition. The technology that they’re testing doesn’t solve a business problem. It’s a new shiny object, but it doesn’t solve a purpose. So I think one thing that we’ve observed from a number of surveys we’ve done is if you don’t have a well-defined problem and you’re not creating economic value, your technology pilots will stall because you can’t get momentum from them. So let’s say if I use machine vision and I have a great tool and have a great application, but it’s not creating value, like I’m not using it for quality as an example. But if I can use it for quality, now I can start to identify improvement opportunities. I can identify defects early. I can prevent that product from going out in the market rather than having to recall it. Now you start to get stakeholder excitement and stakeholder support for adoption of those technologies. Another example is people talked about AI.

Sven Dharmani: [00:43:34] Well how do I use AI more effectively? You can use AI for estimating the price elasticity of demand. So if I’m going to run a promotion on a product, how much is truly the elasticity going to be taking into account all the other economic factors, like the conditions of the economy being one, inflation, interest rates. With a different interest rate, price elasticity might be different. So if I just use the old model, it may not work. So being able to use these technologies to solve a business problem is what creates excitement and value. And I think that is really something that can help the adoption and the use. And that’s something we are noticing from a lot of our clients and client interactions and surveys. And we host a lot of clients. We have two innovation centers. Actually I’m at one of them, at Nottingham Spirk Innovation Hub, and another one, which is in Cleveland. Another one is MxD (Manufacturing x Digital) in Chicago. And we hold multi-client events. And that’s a common theme that comes up at both the innovation hubs is how do I get past a pilot. Because I did a pilot, I’m stuck. I don’t know how to how to monetize the improvement value.

Winn Hardin: [00:44:57] Price elasticity seems to be clearly something that would be solved internally. I mean, because the macro events, the economic conditions, your supply chain, lead times, everything, you know that intuitively. When it comes to establishing, when you were saying earlier that there’s a lot of automated solutions out there where the due diligence basically hasn’t been done. We don’t have the ROI numbers. It’s funny because 10 or 20 years ago, integrators always complained that the client hadn’t defined the defects well enough to be successful. And so you’d put money into it, and you may win the project or not, but it wasn’t clearly defined what the true level of success was. But my question from what you were saying is how much do you think the system designers, the outside companies, the third party integrators, how much do they need to take ownership of helping guide the client for this ROI? Because does the client have the industrial automation chops to be able to fully realize the cost of rework, shipping, and everything else.

Sven Dharmani: [00:45:56] I think it has to be a collaboration. We’ve seen it most successful when it is a collaboration where the client, the manufacturer has defined the problems well enough that they want to solve. They have stakeholders aligned, they have mobilized the team to adopt change. And then the system providers, solution providers, the integrators are bringing the best solutions. A lot of times we see the desire to change is not there. They just want to replicate the old process with a new tool. And you kind of have to do both. You have to reengineer the process inherently to take advantage of the new capabilities. So the clients have their half to fulfill, which is sponsorship, stakeholder alignment, readiness for change. And then you tie it with leading-edge technology capabilities. That’s the recipe for success. It’s not the responsibility of one party or the other. I think it has to be a joint effort because the technology providers can bring in different applications. But not all of those applications would be relevant to their problem that the company is facing.

Winn Hardin: [00:47:15] That’s interesting because it sounds like you’re saying that the onus still mainly resides on the client, on the customer, the end user customer to define the process and what is expected to be achieved. But I’m wondering if the case of AI mixed with traditional either robotics or traditional motion control and robotics case, or AI with machine learning and traditional algorithm based, I’m just wondering if the customer has any idea of what the true capabilities are or are not related to that technology solution.

Sven Dharmani: [00:47:44] They probably don’t have an idea, but that’s the part of the ideation. The technology provider or the integrator has to educate the customer about: here’s the art of the possible. Here are all the things you can do. Then the client needs to say, okay, here are the problems I’m trying to solve. For example, it may be cost, it may be reliability. It may be throughput. And then together they define how do we solve these problems. And then the technology provider helps implement. But the company needs to reengineer the process and drive adoption. So what I’m saying is it takes two to tango here.

Winn Hardin: [00:48:24] Right on. But that technology provider, it sounds like he or she needs to be engaging early. I mean, there’s going to be a huge middle segment that is going to be driven by the client and then hopefully getting their ducks in a row, various champions internally and defining what success looks like in a monetary form in most cases. But that technology solutions provider also needs to be good at conveying to the customer: this is the state of technology. This is what’s capable. So that they don’t waste their time, so that the customer isn’t, “Well I thought you could do this. Now I did all this work, and you’re telling me it’s not actually feasible.” So right then they have a bad taste and they go away, don’t come back for years. So I think that’s something that a lot of integrators can take to heart.

Sven Dharmani: [00:49:11] Absolutely. And I think because I don’t know what I don’t know. And until you show me everything this tool is capable of or this technology is capable of, it won’t open my eyes. Exactly. And then I need to think about, okay, the car can self-park. The car can back out of a parking garage. Okay, so when am I going to use those? I can use the summon capability. Okay, well, I got into a garage where it’s closed at 9 p.m. Now it’s 10:00. I’ll use the auto summon capability because as soon as the car pulls up to the door, the door opens so I can get my car out, even if the garage is slow. So I can think of the use cases. But I need to understand what’s the art of the possible.

Jimmy Carroll: [00:49:56] Sven, I really want to thank you for your time. It’s been a pleasure to talk to you. If anybody out there would like to learn more, first of all, they can go to EY.com or check out EY on LinkedIn. It’s an active channel there. And if anybody has specific questions for Sven or us or would like to join a future podcast, they can reach out to us at Manufacturing-Matters.com. Once more, Sven, thank you so much.

Sven Dharmani: [00:50:21]  Thank you for having me. I enjoyed the conversation.

Winn Hardin: [00:50:25] Thank you sir.

Jimmy Carroll: [00:50:26] Take care. Bye.