EMS@C-LEVEL

APEX 2025 Keynote: Kevin Surace on AI, Automation and Digital Transformation

• Philip Spagnoli Stoten

Share episode

The manufacturing revolution we've been promised for decades might finally be within reach. In this interview with Kevin Surace about his keynote at APEX 2025, he cuts through the hype to reveal how artificial intelligence is transforming production floors in ways Industry 4.0 never quite delivered.

"We want a lights-out factory," he explains, painting a compelling vision where equipment runs autonomously, maintenance requirements disappear, and yields reach 100%. What makes today's AI revolution different from previous technological waves is its accessibility and immediate productivity boost. For just $20 a month, professionals across every function can generate more ideas and solutions than they had "five seconds ago."

The most fascinating revelation comes when Kevin discusses humanoid robotics as the bridge technology between our current manufacturing infrastructure and fully automated facilities. Since existing equipment was designed for human operators, humanoid robots represent the fastest path to automation without redesigning entire production systems. This approach could accelerate reshoring efforts by making labor costs less relevant to manufacturing economics.

Beyond the factory floor, our conversation explores how AI is revolutionizing content creation, supply chain management, and business forecasting. Kevin draws compelling parallels to the early Internet days, reminding us how technologies initially viewed with skepticism quickly become indispensable. "There was a time when the Internet showed up and no one wanted it at work... And today, you can't live without it."

Looking to stay ahead of manufacturing's AI revolution? Subscribe to our podcast for more insights from industry visionaries who are transforming how we design, build, and deliver the products of tomorrow.

EMS@C-Level Live at APEX is sponsored by global inspection leaders Koh Young (https://www.kohyoung.com) and Creative Electron (https://creativeelectron.com)

EMS@C-Level is sponsored by global inspection leaders Koh Young (https://www.kohyoung.com) and Creative Electron (https://creativeelectron.com)

You can see video versions of all of the EMS@C-Level pods on our YouTube playlist.

Philip Stoten:

Hello, I'm Philip Stodin, I'm at Apex 2025, and I'm lucky enough to be joined by the keynote presenter this morning, kevin Starrace. Kevin, great to talk to you. Absolutely fascinating presentation this morning. You had the audience in the palm of your hand through the whole time and I think you had some of them excited and some of them a little bit scared, which is always always a good thing. Really excited to see where AI is going generally, but obviously specifically from a point of view of what's happening here. If you walk around you'll see AI is passes on all the booths. Ten years ago it was industry 4.0. I don't think we ever delivered on 4.0. Maybe AI will help get us across that line. What do you think are the most important things for the companies here to be thinking about with respect to AI when they're looking at the next stage of development of their products?

Kevin Surace:

Yeah. So look, I look at AI and Gen AI today, as you want to use this for everything and every task you do during the day. Yeah, now you could be in sales, you could be in marketing, you could be applications, you could be in design, you could be equipment design. It doesn't matter, you will get more ideas, even from from, whether it's Claude or Gemini or OpenAI, gpt-4 or whatever it is. You'll get more ideas than you had five seconds ago. So why wouldn't you do that? You know, for $20 a month or whatever it costs, why wouldn't you do that? And there's no question that then, beyond that, you say okay, well, I've got this equipment. How could I use AI to improve the outcomes, improve the performance, improve the yields? And, of course, it's imaging systems that are far better. We've had imaging for decades, but we weren't backing that with deep learning models. We were backing it with pretty simple image comparisons, and now we can do something that's amazing, ten times better than we could do before. We can put IoT in all of our equipment. We can monitor everything that's happening in the equipment Flow rates and current draw. It doesn't matter, every single thing, and so the more knowledge and the more data we have, the closer we're going to get to an industry that can run.

Kevin Surace:

Of course, everyone really wants to run lights out factory. That's what we want. We want to push a button that lights out factory. There's no one in there, the equipment doesn't need maintenance, you know, nothing ever goes wrong and the yields are 100%. That's the goal, and part of the reason why we want that goal isn't just cost and getting the people out, but when you have humans involved, humans are imperfect and you know you just can't get all the yields you want. The more people that are involved in the process steps right. So we're going to see AI be able to do that, and the big thing I think we will find is humanoid robotics, because all this equipment that I'm seeing here was designed for humans to be, at.

Kevin Surace:

It's the size, it's how you lift the lid, it's how you put things in, whatever it is, and we need humanoid robotics at first to interact with our equipment because they were built for humans, and once we do that, we're going to be able to onshore and get into other countries that were too expensive labor-wise, because it's really going to be fully roboticized.

Philip Stoten:

Yeah, and it really takes out a lot. I'm really interested in some of the business processes, particularly business processes around supply chain as well. We've just gone through and it's clearly a post-COVID impact in terms of an inventory bubble working its way through and there are issues with that in terms of failure to predict failure to predict demand and it has caused problems. It's caused, certainly in Europe. It's caused the downturn in the last year because the inventory bubble arrived with the customers. Do you think AI has got a role to play in smoothing out those supply chain challenges and actually helping not just the EMS companies but the OEMs do a better job of predicting consumer demand in the way that fast fashion uses it?

Kevin Surace:

Yeah, certainly, and obviously fast fashion does use it, and that probably killed Forever 21,. Who didn't, right, I mean? So look, ai models can be built around anything Right, and we can learn from all of the past bubbles, and AI by itself can certainly better predict that. Now, what AI is not good at are these unique moments that happen, like COVID came out of nowhere and boom, there it was, and there was a black swan event that couldn't have been predicted, so AI couldn't say at any you know next week couldn't have been predicted. So AI couldn't say and any you know next week there's likely to be. So we couldn't predict that.

Kevin Surace:

But we certainly can look at all kinds of things around the world. We know retail sales are a little bit down right now. We can predict that consumer demand may change under current administration because of tariffs, whatever it is right, and so that we can predict down the line six, nine, 12 months, because we know how long the supply chain is, what's going to happen to component prices, by the way, should you hold off buying yeah, you know or should you inventory quickly? It's a little bit like the toilet paper problem you know at at sort of towards the end of COVID, it was like you couldn't get anything because there just wasn't enough of anything and you wish you had stockpiled everything, even though you spent 20 years just getting the JIT and you got to JIT you had no inventory of anything. No, that's right, I think.

Philip Stoten:

What's curious there is that much. As it couldn't have predicted COVID, surely it could have learned from COVID in perhaps a better way than we did and perhaps predicted more of what would happen on the bullwhip end.

Kevin Surace:

Almost certainly, almost certainly, and now that we have that data, we can do that. Yeah.

Philip Stoten:

You talked a lot about the use of AI in content generation. I think it's really interesting and I just wanted to talk to you about that, because what I've started to do is take the transcripts of the interviews I do, I squeeze them into a nice type focus AI model where it's just looking at those. I can take then 20 CEOs from the industry and say what do all those CEOs think about a certain topic, and it has made me massively more productive.

Kevin Surace:

I can then create that material in as many languages as I want and it does a really nice job.

Philip Stoten:

That's a game changer. But for me, I think what's key there is that we have to have done this really good interview at the beginning to get that process going. So it does become that really interesting situation where you've got to have really good learning data, you've got to have really good, really smart prompts, massive productivity advantage, sure, and then someone just taking care of the back end, that everything looks right at the other end.

Kevin Surace:

Well, look, the bottom line is content. When you've built a neural net on a trillion, let's say, llms, on a trillion phrases of, say, the English language or any language, that's kind of the world's database of everything that's ever been said, and so, instantly preparing for an interview, you can prepare now in five or ten minutes, be ready on a whole topic that you didn't know anything about, right? The Neptune C+ it's a revolutionary, true 3D inline dispensing process, inspection. I want that, but I didn't know that much about it, right, but I could know a lot about it in just minutes. I can take an interview and summarize it. I can take an interview there's now video editing stuff that will actually go through this, look at it and say I need 10-second clips, 20-second clips, 30. Boom, boom, boom, boom, boom, done. It's like whoa, I never really wanted to do that work, right? I mean I had to do it. What I wanted to do is be the interviewer Like yeah, that's right, that's when I have the conversation.

Kevin Surace:

So this is making you much more productive, probably 10 times more productive than you've ever been and you're better at your job because of it. You're producing better content because you're more learned on what to ask and what's going on. You can even ask, you know, claude, or ChatGPT or whatever. What 10 questions should I ask this person based on their website here? Boom, you might have thought of them, but I bet there'll be half of them.

Philip Stoten:

you would have never thought of, yeah, and some nuance in there that you'll find Better for the audience, better for everybody.

Kevin Surace:

Better for you, better for me, better for the audience. Everyone wins. I don't know who. The only people who don't like this are the people who don't want to use it. And look, I've been interviewed by people who say, well, I am not going to use this technology because I'm the old-fashioned way.

Philip Stoten:

All right. Well, that was my quill, that's right, that's right.

Kevin Surace:

There was a time we didn't have the wheel, yeah, but people had them. And remember the Internet. I said this this morning, right, there was a time when the Internet showed up and no one wanted the Internet at work because, oh, you'd play games or you'd shop, or you'd do this or you'd do that. And today you can't live without the Internet at work. You just shut down, you just close the shop, right, so all of our equipment is hooked to the Internet. You can't do anything, right.

Philip Stoten:

No, you're absolutely right, and I think what's fascinating is how that change in productivity, whilst maintaining the value of the skill, really is an amplifier for what we can do. What I'm kind of curious about is we all look to the internet and thought that was going to make our lives easier and maybe we could work less. As you pointed out, we're now available 24-7 because of the internet.

Philip Stoten:

We're really working more and we'll work more. Are we actually going to get over the line and actually have some labor saving going on? Are we going to be able to spend some more time chilling out with our family instead of just working?

Kevin Surace:

I have predicted this may happen, but one of the things in the US not true in every other country, but in the US is that we have a declining working population, with more people retiring than are coming in the workforce, and we also have most of our, many of our companies. Most of our companies are growing, not shrinking, and so that means you have to grow with less people, yeah, which means there is more work to do per person. A lot of that can be augmented and help with AI and we're going to have to do that, and some of it, robotics and humanoid robotics, et cetera. But we cannot grow our companies thinking I'm going to double the size of my company by doubling the number of employees.

Philip Stoten:

There aren't twice as many to hire.

Kevin Surace:

It's not possible.

Philip Stoten:

As you say, you've got to double the size of your company with 30% less probably that's right With 30 men, 20 years, 30% less people at the company.

Kevin Surace:

you doubled the revenue of the company, so that means everyone's working. You know well, they're 3x the productivity they are today, something like that no-transcript.