EMS@C-LEVEL
As Forbes, Entrepreneur, Fast Company and SCOOP writer, Philip Stoten, continues to talk to EMS (Electronic Manufacturing Services) executives he learns more about their individual and collective experiences and their expectations for their own businesses and for the entire electronic manufacturing industry.
EMS@C-LEVEL
MTEK is Revolutionizing Manufacturing with No-Code MES and AI Innovations
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
Could a no-code, highly customizable Manufacturing Execution System (MES) be the game-changer the EMS industry needs? Join me for an insightful conversation with Matthias Andersson, the founder and CEO of MTEK Industry AB, as he shares the revolutionary features of their MBrain product. Mattias provides a behind-the-scenes look at the development journey of MBrain, detailing how it stands out in the crowded MES market with its rapid deployment capabilities. These are some of the features that led to MTEK being named a Global Leader in the MES space, and an Innovation Leader by ABI Research!
We discuss the significant challenges faced by traditional MES systems and highlight a compelling case study where a customer achieved zero defect production within just 100 days, underlining MBrain's efficiency and impressive time-to-value.
In another segment, we shine a light on the transformative potential of AI in manufacturing utilizing Microsoft's latest innovation, Microsoft Fabric, exploring how Fabric's data integration and analytical capabilities can impact manufacturing processes. We discuss the strategic alliance between Microsoft and MTEK, which promises mutual benefits and opportunities. As we trace the evolution from Industry 4.0 to today's AI-centric approaches, personal anecdotes from the Nokia days in the '90s illustrate the impressive advancements in the field.
Tune in to understand how combining deep domain knowledge with AI expertise can drive meaningful digital transformation and process improvements in manufacturing.
EMS@C-Level is hosted by global inspection leaders Koh Young (https://www.kohyoung.com) and Global Electronics Association (https://www.electronics.org)
You can see video versions of all of the EMS@C-Level pods on our YouTube playlist.
Hello, I'm Philip Stodham, from my House to Yours. Welcome to EMS at Sea Level. I am joined by Matthias Andersen from EmTech Industry. Matthias, always a pleasure to chat, always a pleasure to talk to you. I've been taken by what I've seen recently from ABI research and the recognition you've got for the mBrain product, and it feels like it is a completely revolutionary version of an MES system. Start by telling me a bit about the relationship with ABI research and how they've actually recognized the product.
Speaker 2Yeah, first of all, nice to meet you again, phil. It's been a while. I'm super proud over the acknowledgement from ABI Research. They reached out to us and they conducted a very thorough run-through of the product, the team, what we're doing, our thinking. Run through of the product, the team, what we're doing, how we're thinking, but, most importantly, what actually is available in the system. That's been ongoing, I would say, for the better part of more than a year now. Finally, when we saw the report, I was, of course, super proud. Sad that we didn't get number one in innovation. There's a reason for it and there's a path forward next year. It's nothing but one. Yeah, no, but it's a great testament to what the team has managed to accomplish. And, of course, all the great input from our customers. Yeah, we're on a path to something else and it's what becomes more and more obvious for us as well, mainly in dialogues with customers that we're not just an MES.
Speaker 2We're more a manufacturing system and we are, since the system is completely it's no code all these features are available out of the books and then the customer value is generated based on how they can help solve their problems or their use cases with the system and that they can do really quick. So it's been an enormously fun journey. The last, remember, it's only three years ago since we launched the system, so it's been short but hectic and really fun, yeah, and exciting, yeah and exciting. And it's been really interesting to see the case studies and where it's been short but hectic and really fun, yeah and exciting.
Speaker 1Yeah and exciting, and it's been really interesting to see the case studies and where it's been used and where it's been really successful. And I think what fascinates me is, you know, when I look at the incumbents in the MES market and you think about what's out there and you think about how the industry responds to it, much of the frustration that you it, much of the frustration that you see and much of the concern you see is the time to deployment, the time to value, as a lot of people would refer to it, the complexity of customization, and that's led many very large EMS companies to just design something themselves because they basically couldn't see a way through. This is super customizable and super fast to not just customize but deploy within your organization. Was that the thinking from the start or were you responding to something you thought was missing in the existing marketplace?
Speaker 2Yeah, definitely my background comes. The only thing I know is manufacturing specifically electronics, and I've been on the developing end and receiving end on similar systems. And if you were to compare the electronics industry or the EMS industry we are, I'd still refer to us as we highly digitized. The tools that we have for the SMT machines are highly developed, but they're still very siloed. There's nothing that really truly supports what happens end of line, manual assembly, tying in information from test systems back to the root cause, and all those things are missing.
Speaker 2So when we went on to develop a system that is possible to be used by, I would say, any normal manufacturing engineer that has a skill set and has the digital savviness, we decided the system has to be no-code, because you can't wait for a company to spend three, six months developing a new feature for you. You need to be able to iterate, just as manufacturing iterates. Every day Things happen. You have to be able to have a system that is hyper configurable and super flexible to create value, and that's something that we have realized more and more, that we haven't pushed for so much as the capability for customers to that. We haven't pushed for so much is the capability for customers to, in a very simple way, understand the method and be able to visualize the flows, because it's all about flow.
Speaker 2We have customers. We have one of our customers now that inaugurated their facility in Poland last week. They've gone from first production 100 days ago. We're now producing thousands of heat pumps, complete heat pump units with zero defects. They've ramped this up basically on their own. Of course, we've supported them and guided them and helped them. Time to value is of essence because we don't have the time, time to wait.
Speaker 1Absolutely, yeah, absolutely, and you know that's an essential ROI measure in itself. What I'm interested in is when you think about this product. Could this product have been developed 10 years ago, or are we in a position where digital transformation, in terms of the backbone, in terms of the AI power that's available to you, has made it possible now, when it wasn't, perhaps, in the past?
Speaker 2I would say it would be possible to do 10 years ago, but the maturity and the level of acceptance among customers Because my theory is that we have four-year technology cycles, everything goes in four-year cycles Now the rate of adoption of generative AI, as launched one and a half years ago, has been just crazy. We have not been helped by AI because we started development before. What we have the foresight to see is that what type of architecture do we have to have? What type of capabilities does the system have to provide in order to be able to utilize tools that come that will be laying on top or surrounding or embedded into our solution? Yes, it would have been able to do it 10 years ago, but the market adoption would not have been as strong.
Speaker 1Yeah, and what about the kind of natural language side of it? You use natural language in your system. You enable your customers to use voice-based instructions. Is that something that is significantly better now than it was in the past?
Speaker 2of course, no, no, that technology is evolving crazy, and we are super proud of our partnership with microsoft, so we get access to it, to, to platforms and systems that will be on the market 12 months from now, yeah, which just shows that we are just on the beginning of this journey. What they're capable of doing that will help every manufacturer contextualize things and in that dialogue as well, is what Embrain provides to customers is the capability to create an end-to-end flow where data is tagged and flagged in the correct way, so that you can truly contextualize the models as well as you can bring in other streams of information that you weren't able to bring in before, which then, of course, helps strengthen the entire AI case.
Speaker 1Yeah, yeah, and it builds up to business intelligence as well as manufacturing intelligence and connects to so much more.
AI and Manufacturing Revolution
Speaker 2Yeah, no, take an example. Now we'll talk about Microsoft. Microsoft just launched Fabric. Fabric is the fabric. What ties everything together? Well, we got access to an early private preview on capabilities that will be launched later, within five minutes and I promise, five minutes. After we got access to it, we made the integrations to it and was certain I can't name the features that will be released. We were able to get our system through a prompting to answer three, four, five of the most important questions that any manufacturing engineer would ask themselves every morning. And the data was there, already structured. And I was blown away of that capability, because imagine if I had that capabilities 25 years ago when I was working for Nokia. So it's, we're at the shift change.
Speaker 1It's a bag of popcorn and uh, yeah, watching what happens, it's just amazing. Yeah, sit back, sit back and enjoy. It's going to be a big show, isn't it? It's really, it really is very exciting. And I think it's important to say that that relationship with, with microsoft isn't bought, it's earned. You have, you have to be doing something special for them to agree to partner with you, and they're not partnering with every MES company that's out there. They've chosen M-Tech quite specifically.
Speaker 2Yeah, I'm super proud of that. The relationship that we have and the mutual benefit that we see is we see great potential, of course, yeah, yeah, that's huge.
Speaker 1I wanted to shift gear a little bit and just kind of explore the AI possibilities within the electronic manufacturing space, and it's really interesting. I was talking to someone at Apex this year and we were saying that two or three years ago every booth had Industry 4.0 on it and now every booth has got an AI logo on and one had you know the two, basically two AI letters that were both taller than I was on their stand. It didn't really tell me what they did with AI, but they really wanted to let me know that AI was important to them. Where are we, do you think in that? You know, are we very much still in the hype cycle? Are we starting to see decent deployments? Is there a concern that we're going to drop into that classic valley of disappointment before it starts to climb again? What do you think?
Speaker 2There are so many activities that have true valuable use cases that start to surface, I think.
Speaker 2Of course now I'm biased, I'll try not to be and try to be more generalistic, but I think that the key thing with AI is that you have to have information from areas that you haven't had information from before, which means that looking to a machine, machine, data provides a lot of information. As long as it's typed correctly and get all the information, it's good. But what happened during the changeover? What was the skill set of the person that made that changeover? Did you set it up properly? And I'll give you a personal story.
Speaker 2When I worked for Nokia, I was part of the data management team on Nokia. I got the opportunity to be definitely the most junior person on the team, but we deployed neural networks. This is mid-90s, end-90s, if you remember the 6150 and 3110, the brick phones. We were able to deploy neural networks SPC, simple on cycle time analytics, and what we saw then, if you take those, were the old Fuji CP6 and CP4s mainly CP6s and watch the cycle time. So 40 seconds, 40 seconds, 40 seconds, and suddenly you have those microstop, which is 40.1.
Speaker 2The 0.1 is like when you go to the doctor. The doctor looks, takes a stethoscope and listen to your heart and if there's, first of all, you have a pulse, good, you're alive, that's a good sign. But then you start to look for the anomalies. Because, based on his training experience and he knows that well, as in my case, slightly overweight, my blood pressure is high, so those things affect the heart. Same thing goes for when you deploy AI on the manufacturing floor. You need to contextualize as many areas as possible in order to draw some conclusion from it.
Speaker 2We were able to predict the process outcome of those 31 tens before they hit the test station, because we saw that when you had a 0.1 deviation and you went back to the nozzle, you went back to the feeder and you had a pickup error or you had a nozzle error. That nozzle error would probably cause more errors. And then so those things we were able to predict, and that's way before. Co-pilots, yeah, of course, but now, with the technology that are available, with all the systems and you can tie everything together, we will start seeing some super valuable use cases, and genealogy and traceability, of course, is, from my perspective, is the strongest tool and data source for process improvements if you use it correctly.
Speaker 1Yeah, yeah, it's it, it's. It's fascinating, matthias, and you know when I, when I think about the ai opportunities are out there and I think about where we should be getting them from. I think of companies like mtech but not just mtech where I think they understand ai. That's great, that's important, but there are a lot of people that understand AI. They have a massive experience and domain knowledge in manufacturing and that is also hugely important that you know. I want to buy my supply chain AI from someone that understands AI and supply chain not just AI and not just supply chain and I think that's the combination, and that's where those partnerships with Microsoft, but also your understanding of how manufacturing products are put together, going all the way back to Nokia, is so important.
Speaker 2It is that combination of deep domain knowledge with this broader scale of digital transformation and ai that that is going to add value yeah, because at the end of the day, we're we're only working with people and if you can't convey, because I think that's the strength of generative ai that they can actually contextualize their response back so you get truly actionable insights and you also get guidance on how to fix the problems. Downtime management and connected to scrap scrap management, connected to changeover procedures, connected to just lead time reductions in general. But there is, it's a fabric, everything is tied together and if you can help a young person that enters the industry or whatever industry they enter climb the ladder to, to to knowledge and experience much faster through these super cool, advanced tools, well, what have you we created with that? That's amazing leap that we'll take, because we know that we have a demographic challenge in the whole, in the entire world. We have too many, too few young people that want to join the industry because it's not cool enough or you don't get the right tools and stuff like that, but that is about to change.
Speaker 2Yeah, I had one of my most senior uh programmers coming running to me jumping up and down when we activated our AI tools internally for developers and he said what usually took me eight hours I now do in less than three seconds Wow, which is of course. Most AI initiatives fail because you go for the hype but you don't understand. How does it create value?
Speaker 1And that's also a key learning on this.
Speaker 1Yeah, and that comes from that experience as well.
Speaker 1I'm curious to see if we were a lot further with AI at the outset of COVID. We've had this massive bullwhip effect as a result of COVID and a lot of that has been because we haven't been able to predict demand and we haven't been able to manage the supply chain. It would be really interesting to see how really good, well-managed AI with the right learning sets, with the right questions, can help us perhaps in the future to manage demand, because I think you know, that's something that we find very difficult, and I was interested recently where I saw a business story of Shein, which is this ultra fast fashion brand that actually manufactures entirely on demand. So they're not manufacturing products to store, they're delivering direct. They're using AI to predict demand and it is incredibly accurate and it is creating a more sustainable model. It's creating less waste. It's creating better value for the consumer as a result of that. So I think that control of the supply chain, particularly the demand side, which we're pretty much guessing at the moment, is really fascinating as well.
Speaker 2It all comes down to. In order to be able to utilize that, you need to, beforehand as well, understand where waste is generated, because if you can't reduce your lead time, you won't be able to be flexible enough, you can't reduce your defects, you won't be able to be predictive enough. But key thing here is to be able to build the foundational data layer where you collect your sources from, so you can actually be truly accurate on those predictions as well. To really use the power of AI, because it's not a magic wand that you're going to swing over everything Suddenly. I have a hundred changeover or I have a two hour, three hour, four hour lead time throughout my production. That just doesn't happen. You have to think from the design perspective. You have to understand the flow, how it's constructed. You have to build in all those necessary All those necessary gates, quality gates, understanding where your bottlenecks are. Work with those things. We always have been working. People provide better insights for more people. You won't be as reliant on more experienced people.
Speaker 1You can grow your experience much faster. I think that the key thing yeah, that's it. It's access to it's access to multiple experience, but it's the people with the experience that you've got, uh, matias, that are able to think about what the training sets need to be, think about what the which questions need to be asked um before they're answered, and I think that's hugely valuable. Um, I think we're going to hear a whole lot more about Embrain in the coming 12 months. I look forward to hearing more about the Microsoft partnership. I look forward to seeing some of your really interesting customer case studies on LinkedIn. I think they're really valuable and I've enjoyed those very much, and it does seem to be a game-changing technology. So, congratulations again and thank you so much for your time. Thank you.