Show Notes & Links

We encourage you to listen to the first episode of the two-part series, “What is Digital Transformation and Why it is Important?” 

Episode Transcript

Mike Ambrose
And so things that would take like six months, you can do in less than a month. You know, cost we’ve demonstrated and not just at Sikorski, I’ve seen McKinsey studies and other studies that show you see reductions in cost, like by over 50%. You see improvements in quality by over 90%. Those are all real.

Tan Deleon
On behalf of the members of the Connecticut Academy of Science and Engineering, welcome to this edition of Learning and Living STEMM in Connecticut, the podcast of the Connecticut Academy of Science and Engineering. My name is Tan Deleon again, I’m an elected Academy member, and I serve on the Academy’s Governing Council. For more information about the academy, please go to www.ctcase.org. I’m pleased to have as our returning guest, Mike Ambrose CASE member and recently retired Vice President of Enterprise Business Transformation at Sikorsky, a Lockheed Martin Company, for part two of What is Digital Transformation and Why is it Important. Welcome back, Mike.

Mike Ambrose
Hi Tan – so glad to be back. We had a lot of fun the first time, and I’m looking forward to continuing the journey.

Tan Deleon
Same here, same here. I’m definitely hungry for some more intellectual nutrition. That was a fantastic talk, and I think our listeners are going to enjoy part two. So let’s get right into it then. In part one, we talked about digital integration, and integrating the interfaces of a product value stream. Can you give our listeners a quick recap?

Mike Ambrose
Sure. And, you know, if you go back to that first discussion, we started with the product lifecycle. Could be anything – could be a helicopter, could be a washing machine, could be a toaster, could be a car – we used a toaster as an example, and I might pull that back out again for today. But really the whole, the whole idea is that in a product value stream, you start with a customer need requirements, you go through design, you go through build, assembly, test, validation, you then go into service get in the field. Basically, any product follows that same value stream. The key is from a digital integration standpoint, digital transformation is that there’s data along that whole product lifecycle, and more and more, there’s data as we become more electronic. The idea is, is that by using that data, ideally, through a single source of truth, a single source of data, is that you’re able to manage the interfaces between those individual steps. And so in the first part, we talked about the complexities of managing those interfaces, and also what’s enabled by being able to use that data. So with that, let’s go and talk some more.

Tan Deleon
Yeah, no, I mean, that’s definitely a nice recap. And, if anybody, you know, hasn’t listened to the first part, definitely listen to the first part before continuing with the second part, just to make sure that everything follows in a nice sequence. So, just picking up on some of the tidbits that you just mentioned, you also mentioned, you know, digital twins as the Holy Grail. Can we start there today? What does this actually mean?

Mike Ambrose
Sure. So you know, digital twin, we talked about how it’s a virtual representation of a physical product. We will use the toaster because everyone can relate to the toaster. But it could be anything. And that digital twin represents every function, every interface, every aspect of that, and when we talked about the toaster, we talked about, you know, the nickel, chrome filament wires that heat the toast, we talk about the springs, we talked about the levers, the resistors, the wires…We even talked about what makes a great piece of toast and the chemical reaction that occurs that gives that caramelized outer coating – how it needs to be dry. Quickly you can get into the thousands and tens of thousands of interfaces that make up something as simple as a toaster. Now the digital twin would be a digital representation of all of that, every aspect of that, that you would then be able to go in and optimize designs, optimize the way the toast is made. Optimize for how long in life this operates, and how many times you go up and down with a lever. All of that could be data. And so from a digital twin perspective, taking all that information now allows you to have a digital representation of what’s happening.

In the first part, we talked about a very important concept called X before you X, design before you design, build before you build, test before you test, maintain before you maintain, and there are many, many others. The concept that when you have something like a digital twin, is that now you’re able to do these simulations and do this X before you X much more efficiently and much faster. Things that would take days, weeks, months, years. If you have sufficient data and you have the digital twin, you can do it in a matter of seconds, minutes, or hours, depending on the complexity. And so when I talked about the digital twin being the Holy Grail, it is aspirational in many regards, because you’re always driving towards what is a complete digital twin. And so when you start to think about that, the industry has made a lot of progress, in some aspects, around the design and build assembly portions of the product value stream. That’s where you see the most benefits coming out of the industry because it’s the most mature. Computer-aided design, and computer-aided manufacturing in some form or another have been around for as long as 40 years. And so that maturity and being able to attribute data, being able to collect the data and understand how to use it to create simulation, use virtual reality, and other and other tools, has enabled significant savings, particularly in the build portion of the value stream. And also being able to optimize on designs and being able to do things like generative design and additive manufacturing, that you couldn’t even do ten years ago. All of that is enabled by the data, enabled by constructing the digital twin. So it is, I call it aspiration because you’re always driving towards that nirvana. But there’s a lot that goes into limitations and constraints and what it’s going to eventually take to have a comprehensive digital twin.

Tan Deleon
Okay, yeah, so the sort of the notion of X before you X certainly makes makes makes a lot makes it a lot more clear with respect to the digital twin because now you are able to X before you X and actually validate that you are X-ing before you X-ing properly. Right?

Mike Ambrose
Yeah, really important concept because you even taught you had you had made the great insight last time around, “Well, how do you know, it works?” So there is the way the state of the art is today, is that you do have to have a correlation, you do have to have some sort of physical testing so that you understand the boundaries, and you understand the constraints. So that you can use the simulations, that you can use the data in a way that you’re interpolating, you know, without having to do a lot of extra testing. So your validation is a very important concept. Now, over time, as you build this data, as you build confidence in the models and the technology capabilities, you’re able to start to extrapolate and to extend those simulations more. And that’s more of what you’re seeing in generative design and with with, you know, basically build before you build.

Tan Deleon
So you also mentioned like, like a digital thread, right, and having, you know, different threads that go to, you know, depending on what the actual application is, you can have a thread for like the piece of a component in the toaster, or maybe the toaster as a, as an actual complete system writ large, so to speak. So, like, what is the timeline for having a fully functional digital thread, right? And then is there such a thing as a certification of a digital thread? Because that probably goes in line with what we were talking about with the validation of that digital twin. So that you know that when you’re X-ing before you’re X-ing it’s actually done properly.

Mike Ambrose
Yeah, really, really important concepts, especially when you’re doing things like you know, for government, defense applications, safety applications. You know, how do you certify? How do you know it’s good enough? There’s a lot into that. So starting with your first question in terms of the digital thread. Ultimately because there are different constraints, there are different levels of technology – and that’s always evolving, always changing – which hold on to that thought, because that’s also going to play to the certification. Is it that, you know, as you’re going on this journey, and you’re getting more capability, more computing power, all kinds of decisions on so what’s important, what do I need to measure… because there could be gigabytes of data – we have gigabytes of data every time we would fly a helicopter – how do you know what data to use? And so all of that is the evolution. The challenge, when you say a fully integrated digital thread, the aspiration, the goal should be on connection so that you can measure, you can improve certain aspects. And I think our discussion last time was effective in terms of understanding the components or subsystems so that you can make sure that they’re optimized. That you probably start to lose value as you try to do too much, at least for now. Too comprehensive, it’s too expensive, and takes a lot of computing power, and a lot of storage. All of those are trades, and all those are costs and all those are scheduled, and all those are risks. And so the beauty and the balance is as it matures, how do you find the right thread to go and maximize savings, maximize cost? And being able to go and validate. So it is as dance you play with the technologies and with the capabilities always measured against key things like cost and schedule and ability to go in and certify So the best way to answer that is it is an ongoing trade.

And, you know, in terms of like certification, you’re starting to see organizations start to begin to certify digital twins. And my counsel there is that it’s inevitable. I think that at some point, there has to be some sort of standards to say, we’re at this level. the technology is changing a lot, in a good way. And so what you’re gonna probably see, and you’re seeing this, like marine applications, I think that’s from my research, they’re a little bit further ahead than other areas, in terms of specific applications of a digital twin mostly for maintenance, you know, in terms of things that are important to the users. And so you’ll see specific threads, as we talked about last time, that will be the subset of probably how certifications will start. Otherwise, it’ll be too general, at least for now. And, you know, I think the technology is changing so fast, it’ll become obsolete or stale in a hurry. And like anything, once you put a structure around it, it does begin to limit the innovation. The way I looked at it in my career is that you’re always measuring the bottom line, you’re always measuring, “Am I doing it faster?” “Am I doing it for less cost?” Cost in many dimensions, you know, could be used, you know, like cost in the field, could be cost to acquire it. So you’re always looking at cost, you’re looking at time. And then you’re also looking at performance and capability. So there’s bottom line metrics that you’re looking at those early indicators in terms of the digital tools, the digital enhancements that you’re making – are they driving, you know, less costs, faster schedules, better performance? And there are leading indicators and how you can monitor that over time.

Tan Deleon
Okay, so then – you know, from the certification perspective, you know, just to expound upon that just, just a tad – it sounds and please correct me if I’m wrong, it sounds like the processes or the standards that are in place to certify, are based on each individual entity’s own set of requirements. So there isn’t like a standardized way of doing the certification across the board. Is that Is that right? Or is that wrong?

Mike Ambrose
Well, it is actually, it’s an interesting concept because part of the evolution is the way we’re going to validate and verify requirements is that, you know for the most part, there’s a lot of requirements that are – for all good reasons – are there today that you have to meet from the lever to the electrical wire here to the buttons. There are design requirements, you know, we can imagine for something like a helicopter or jet fighter, it’s, you know, it’s 10s of orders, it’s 1000s of orders of magnitude more complex. And so those requirements exist, they’re always there. And so now, what you’re looking to do is, how do you use simulations? How do you use digital integration? How to use things like model-based design, and model-based system engineering, to complement that? And you know, basically use simulations at the end of the day, to take the place of things like physical testing. And so, you know, that’s, that’s the maturity that you’ll see over time. But the reality is that you know, whatever, whatever is done from a digital integration standpoint, that transformation standpoint, needs to, you know, complement and correlate to existing requirements, as they exist today. Yeah, we’re gonna get some real nitty gritty, system engineering.

Tan Deleon
Yeah, yeah, okay.

Mike Ambrose
But it’s important, because, you know, what’s interesting is, is that, that’s the reality is, is that you can do amazing things in these simulations. And there’s still a structure that we need to fit in. And, you know, as we go and do that, there’s healthy, there’s healthy conflict. And, you know there’s been many times where, you know, I’ve had discussions – in some cases with three and four-star generals and admirals and including, you know, people from the companies I’ve been associated with – where’s my savings? Why does it still take so long to go to develop products? And, you know, and I think that’s some of the tension that you’re seeing.

Tan Deleon
Yeah, so speaking of high-ranking military, you know, recent senior military officers have gone on record as saying, you know, digital transformation is overhyped. Can you talk about about this a bit more?

Mike Ambrose
Yeah, I’ve seen actually, I’ve seen that quote myself. So um, so you know, I was directly asked – I remember not too long ago by a three-star general – and said, “You know, all this is great. And every time I hear companies talk about it, I just, I don’t see how I’m saving time, how I’m saving money.” And so, fortunately, at Sikorsky, we had very specific examples of how we were saving. It goes back to the digital thread, the digital twins are not maturing at the same rate throughout that product value stream. So, where you see the most maturity is really between the design in the manufacturing. And what we’ve been able to demonstrate is that – and it isn’t just Sikorsky – there are many other examples of where you can leverage computer-aided design and from a simple terminology, and everything that feeds into it, which ultimately is model-based design. You’re putting attributes against features and capabilities, on a part, digitally. And by doing that, you’re able to do these design before design, you can start doing generative design, you start going out of manufacturing in many cases. And now you’re also able to provide that information to a manufacturing engineer, a quality engineer, a technician and they’re able to go in iterate really fast, provide feedback real-time. And so things that would take like six months, you can do in less than a month. You know, cost we’ve demonstrated and not just at Sikorsky I’ve seen McKinsey studies and other studies that show you see reductions in cost, like by over 50%. You see improvements in quality by over 90%. Those are all real. And I remember one time I was asked by this senior executive at Lockheed – but you know, the same question that the three-star general asked, but it’s still you’re finding things and still there’s discovery. And what we point out, when I pointed it out it says, yeah, that was not an area… This is one particular discovery that was modeled in a digital perspective. We really don’t have a digital representation yet, of that particular feature. And so that’s the aspiration. That’s the holy grail aspect of this is that the race to go and represent what matters now. It’s always a cost-benefit analysis. Maybe there are decisions, you know, we use a toaster, for example, do you need to digitally represent everything about a toaster? That’s some of the intelligence and things like artificial intelligence/machine learning is helping make decisions around what needs to be modeled, what needs to be attributed, so that you’re making sure that the things that drive the overall cost, the overall scheduled performance, are being modeled so that you can simulate them and be able to go and risk reduce before you actually go do that task.

Tan Deleon
Okay. Yeah, no, that makes sense. I mean, it almost sounds like maybe if there was a standardized way of, you know, picking what you needed or what attributes you needed across the board. And all corporations, you know, signed up to that, that way of doing it, that might potentially help with, you know, the velocity. And obviously, if you’re going faster, you’re probably, as a result of going faster, reducing the cost of doing business. But to your point, if you go too fast, and you make mistakes, your cost is going to be a lot more on the back end, just because of corrections and stuff.

Mike Ambrose
Yeah, no, I love the way you’re thinking and actually, maybe you’re gonna get a consulting job out of that. Because we did not rehearse or talk about that. But ultimately, yeah, you know because it is about what is important. And as you figure out, okay, what do I need to attribute? You know, you look at these, like use cases, and you’ll hear more and more about that over time. You’re making a machine fitting and machine fittings have applications and 1000s and hundreds of 1000 different applications, but there are certain aspects to machining a fitting or additive manufacturing a fitting, which combines, you know, two parts, for example. Is that, what is what is attributable? What is important? And so, you know, I think following your line of thought is, is that, that type of generalization is probably the right way to go after it. I have to think more about that, but it’s just, you know, that’s maybe one of the great things about these kinds of forums is that, you know, you get different perspectives, but ultimately, you know, that, that probably has, that probably has some value in progressing that more.

Tan Deleon
Okay, now, yeah, thanks for indulging me on that. So, you know, we were talking earlier about, you have a run for, you know, one of the one of the helicopters at Sikorsky, and so gigabytes of information, you know, do you think that AI has a role in digital twin? And what do you think that role is? Because, I mean, everyone is always like, oh, well, AI? Oh, AI, like, AI is the panacea of everything, you know, so like, from with your expertise, what do you think the role is for AI in this whole digital world here?

Mike Ambrose
Yeah. And, yeah, it is interesting. I mean, people have all kinds of reactions when they hear artificial intelligence/AI. You know, my perspective on it is, is that artificial intelligence, and you’ll hear me say, artificial intelligence/machine learning, because that’s kind of the state of where its right now. If you look at the maturity and the different levels of artificial intelligence, it’s an enabler. Enabler is really an inadequate word in terms of I mean, it gives life to the digital thread, the digital twins, because when you start to use things like machine learning, data analytics, and artificial intelligence, you’re able to interrogate the data, you’re able to pull out and do analysis and simulations, much, much faster than any human can. You do not eliminate the human is that there still needs to be an interpretation of the data, you still need to be able to understand what the data is saying, and then be able to iterate on that. That’s the state of the art of artificial intelligence/machine learning. It’s an enabler for being able to do these simulations a lot quicker. You know, there are things called like predictive maintenance, prescriptive maintenance, and that’s where you really see artificial intelligence shine, where you’re using sensors and sensors are a big part of it. Whether there’s sensors on like, lathes in the factory, or its sensors on you know, transmissions on a helicopter. Those sensors provide vibration, they provide temperature strain data, that enables it basically data analytics, artificial intelligence/machine learning, to go and interrogate those and start looking for trends. And it could be trends in so many different dimensions that, you know, as that data analytics polls and interrogates, like that flight that has gigabytes of data, it tells the test engineer what needs to be tested more, what to look out for, information on constraints on being, you know, how many g’s should the helicopter perform, to what altitude, to what temperature? You know, it’s just all kinds of information. And then when that – we’ll use the example of the helicopter – goes out in the field. And you start to look at like the life of components because helicopters being the vibratory environment they are, vibration matters, you know, you’re able to go and see things that we would have never been able to see, even 10-15 years ago. So yeah, AI is a huge enabler, I mean, huge fan of being able to apply Data Analytics – does not eliminate the engineers – it facilitates and enables engineers to do their jobs a lot quicker. That’s the state of the art as it is today.

Tan Deleon
Okay, I’m glad you brought up the fact that it doesn’t eliminate the engineers, because I mean, there’s a great fear out there as to what AI is going to mean for jobs, you know, and the future of jobs. So how does digital transformation, the digital thread, and AI impact organizational skills? And, you know, the jobs that, that those skills will need to entail.

Mike Ambrose
Wow, yeah, I mean, that’s a huge question. And to me, it’s one of the most exciting aspects of digital transformation, and everything that goes with it is that organizations are going to need to transform more so than at any point, during the Industrial Revolution. I mean, basically, the hierarchical organizational structure that has been around since the Model T, for the most part, 100-110-115 years, really has not evolved. I mean, you see, like Tesla, and others, and you’ll see small companies attempting to move to more of this agile structure. But for the most part, if you go back…You know, the way I describe the product value stream, even the way I explain it, you know, starts with requirements goes through design, build, and going all the way to maintain, it’s sequential. And if you think about it, with a digital thread in everything I’ve described, is that there should be this incredible concurrency. And there is, you know, there is concurrency that happens. However, a key limiting factor is the organization. Is that you have very discrete roles and responsibilities at every step of the way, that become barriers to being able to go and you know, and I’ll use the word barriers, strong word, to really go in and condensing that as much as I think the tools and the processes and the vision could enable it.

And so with that, and I don’t have an answer, you know, I think a lot of people know the potential is there, and you hear things like agile organizations and agile processes. And there are all kinds of standards. And you know, what makes an agile process to the point where we were having before – and I see you smiling, that’s usually the reaction when someone says the word agile. Now I look at it as agile is really adaptive in terms of how you want the organization – the skills, the talents – to move with the product. It’s almost like it’s an organism – it is an organism – that would be this nirvana. What that looks like, I think is going to take some evolution, it will take some evolution. But at the heart of it, really is, that you want parts of the value stream and we’ll take manufacturing engineer as an example. I started my career as a manufacturer engineer. And I can say with all honesty you know, it was not always the most glamorous. If you want to be an engineer, manufacturing engineer wasn’t like the, at the pinnacle, like aerodynamicist or something like that. But really, when you look at what’s happening today, with digital transformation – I had the terminology when I would talk to my teams at Sikorsky, that I want every manufacturing engineer to be a system engineer. Because the manufacturing engineers are the conduit between that design requirement’s build to what happens afterward. Everything really funnels through them, and the manufacturing engineers were the ones, because that was the most mature part of the digital transformation, they were in the best position to be able to bridge that gap and be able to show how concurrency can be leveraged. So we put a lot of effort around really upscaling the skills of manufacturing engineers to allow them to think like system thinkers. Now, when people think digital transformation or digital, they think engineers. You know, my perspective on it is that everyone in the organization needs to gravitate towards more of the system thinking. That’s the next step of being able to really leverage digital transformation – the more you have human resources and finance and legal, be able to go in and really start that thing from a system perspective – now you’re connecting those interfaces and you’re looking at interfaces in different directions. And this is where that symbiotic relationship with artificial intelligence/machine learning, and being able to understand and interrogate the data. You look at finance, for example, how do you leverage your business cases and really be able to point to the things that are going to save money and do things quicker? So a fascinating area, fascinating, and one that I’m really looking forward to seeing how it goes in the future.

Tan Deleon
Yeah, no, thank you for that. And you heard it here, folks, you know, the days of stovepipes and rice bowls are numbered. So I mean, that’s that, that sounds like the term concurrency is the way to go and the system engineering approach… but not just calling or changing the title of your employee’s role, it’s actually developing those skills, you know, and making sure that they’re concretely founded within your organization. So very, very good points, and very, very insightful. So I appreciate that, Mike.

Mike Ambrose
Well, thank you.

Tan Deleon
So let’s, you know, change slants real quick. Now that you know you’ve spent a very good part of your life developing all the expertise that you have, now that you’re retired from Sikorsky, you know, what have you been up to? What have you been doing since?

Mike Ambrose
Yeah, I always tell people to retire with quotation marks, so you can ask my wife – many things. No, and it actually it’s, it has been incredible. You know, first, I’m a board member for RBC Bearings, a public company, an amazing company. I’m on like, three other advisory boards, startups, and just fascinating because this is a way for me to contribute and at the root of it is that, as you can see, I love the learning, I love helping, giving back. And, and that’s an important part of what I do. I’m Vice Chair of the Board of Governors at the University of New Haven, I’m actually leading the search for the next president. But just even that experience, in terms of working with students, and working with faculty just really excites me, and allows me to continue to keep learning and giving back. I am consulting and one of the nice things about that isn’t just around the digital transformation that I’ve spent, like the last couple years doing, I’ve been doing some real meat and potatoes aerospace consulting, which, you know, is a lot of fun. Speaking on technology like this, and with all that, what’s what’s really nice is that my wife and I are finding time to travel. You know, we both like the landscape. You know, actually [speaking Polish]… so I started learning Polish, like years ago, had kind of stopped, but I’ve started taking formal Polish lessons. I’m not Polish but when I was working for Sikorsky, we spent a lot of time spent a lot of time there. You know, we have a factory there, I was able to make a lot of friends. So you know that and I’m actually running again.

Tan Deleon
All right!

Mike Ambrose
For an old guy, you know, just learning how to run quickly and stay in shape as an old man has been fascinating. So yeah, I’m pretty bored. Yeah, as you can see, not a lot going on.

Tan Deleon
But yeah, it sounds sounds like you just kick your feet up. And you know, you’re just waiting for the sun to set. But now you’re, you’re doing some phenomenal things. Wow! I mean, that’s a that’s a full schedule for someone that’s, you know, not retired.

Mike Ambrose
But, uh, yeah, the big difference is, you can manage your own schedule, but it’s, it’s good, and forums like this, I think are really important. I mean, that’s part of the giving back. So you know, thank you, Tan, for being such a great host and very insightful, and you gave me some things to think about here.

Tan Deleon
Yeah, no, thank you so much, Mike. You know, thanks for taking the time within your current schedule to even do this, and then coming back for a second opportunity was very, very tremendous for us. So, you know, our listeners appreciate it. And I certainly appreciate all the insights that you’ve provided today.

Thanks to our returning guest, Mike Ambrose. For those living in Connecticut and others tuning in from outside our state, we enjoyed learning about part two of DX. If you have not heard part one of DX, I encourage you to listen and subscribe to this podcast on Apple podcasts, Google podcasts, Spotify, Amazon Music, or YouTube, and visit the Academy’s website at www.ctcase.org to learn more about our guests, read the episode transcript, and access additional resources, as well as to sign up for the CASE Bulletin. Once again, Mike, thank you so much for coming back. And this was just an insightful discussion.

Mike Ambrose
Thank you, Tan.