From Connectivity To AI: How Factories Finally Move Data And People with 4IR.UK's David Graham

Factory data doesn’t transform a business until people do. Live from Productronica 2025, I dig into why a decade of Industry 4.0 felt slower than promised and how the tide is turning as engineers connect lines with Hermes, bridge legacy SMEMA with adapters, and elevate insights to the factory layer through CFX. Our guest, David Graham of 4IR.UK, has spent years pushing past vendor silos and championing a simple idea: the most valuable interface in manufacturing is still human-to-human.

We break down what actually moved the needle—reliable machine APIs, repeatable data flows, and a pragmatic standards stack—and what didn’t, like bespoke one-offs that trapped value in custom code. The conversation tracks the journey from adjacency wins, such as SPI-to-printer feedback loops, to line-level stability and then to agentic AI that summarizes performance, accelerates quoting, and targets high-fit customers. Along the way, we examine why success stories remain hidden, how logos shifted from Industry 4.0 to AI without the necessary storytelling, and why internal “marketing” is really about keeping momentum and trust alive.

Talent is the hinge. New grads attracted by real AI problems are joining experienced process engineers on the shop floor. When domain knowledge meets modern software, pilots become production. Standards like Hermes and CFX turn integration from a one-off chore into an asset, giving teams clean data they can trust for quality, yield, and cost improvements. If you’re wrestling with ROI, start small, measure fast, and let the wins compound. Subscribe, share this with a teammate who owns connectivity or quality, and leave a review with the one AI use case you want solved next.

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.

Transcript

SPEAKER_01 0:07

Hello, I'm Philip Soton. I am here at Producttronica 2025, and I'm on the Global Electronics Association booth. I'm joined by David Graham, he's the product manager, product manager director at 4IR.uk. Tell me a bit about yourself. You've been a bit of a pioneer and a bit of a spokesperson throughout the industry 4.0 journey. So a lot of people will already know you, but tell me what you're doing right now.

The Silo Problem: People Over Data

SPEAKER_00 0:32

Thank you for saying that people know me. But yeah, that that's I I think the the issue with our the uh our industry has always been siloing and the lack of conversation. Yeah, you are clearly the man who I've uh watched for 19 years, probably I've been in it since 19 years, since 19 years, so you've been doing different things along the way, yeah. And uh I know that because I'm proactive and trying to seek that information, yeah. But I've I find generally because if it's a siloed industry, yeah, regional, customer base, yeah, um supplier sort of uh supplier-customer kind of relationships, yeah. There's lots of silos in the mix. Yes, absolutely. So just try to resolve that uh and starting a business at the same time, just to be a bit more vocal and a bit more visible and a bit more opinionated and get those sort of views out there, yeah. Um because then that triggers off debate.

What Industry 4.0 Actually Achieved

SPEAKER_01 1:30

A whole lot of debate. And it's really interesting you talk about you talk about the siloed way people operate in terms of their um relationships with their customers and their and the business relationships. It's fascinating that as we've gone through this industry 4.0 journey, we've been trying to get data out of silos, but it's really getting people out of those goddamn silos. That's there's that's even more important. What do you when you look back at those? It feels like it. I I guess it has been a decade, but when you look back at that decade of industry 4.0 evolution, what do you think we've achieved? I I I I'm a little bit disappointed. I feel we didn't get a big dividend, but I do feel we've at least got connectivity and data, and that will lead me to my next question. But what do you think?

Hermes, SMEMA, CFX: Standards And Gaps

SPEAKER_00 2:16

What have we achieved? Well, I think that the the the reason why you're uh not satisfied is because I think we didn't identify the actual challenge which you've already hinted at it, which is people, yeah. And uh we have a uh motto connecting um manufacturing with a digital economy, yeah, which is to basically say that there was lots of software developers working in the dot-com, yeah, working in the more glamorous world, yeah, uh that we could harness and use, yeah, but they need domain knowledge to be wrapped up to work in our industry. So therefore, you need a common ground and data being one of them. Yeah, so um if we saw it as a recruitment um drive and challenge, then uh we probably might be in a different situation. Yeah, um, from a data point of view, a connectivity point of view, we clearly that just takes time. Yeah, I mean we are we have made progress, yeah. And uh just being at this show today, yeah, I've seen exciting customers. Yeah. Um so I'm here sort of promoting Hermes, which is the machine machine community standard, yeah, which kind of lives on top of CFX and configuration. Um but that's always been driven by the equipment manufacturers. And if you know anything about our industry, yeah, it's normally driven by the customer kind of thing. Yeah, so for the first time, there is an interface that's being released that no one actually was gonna use from day one. Yeah, so it's innovation, yeah, you know. Yeah, so uh we've had a bit of delay, natural delay, organic people were on board right at the start, yeah, and then they did it, and then they said, Well, what do we sell? What do we do? What are we gonna sell? Yeah, and uh and then they sort of took an energy sap from that. Yeah, but there's also other people, we we create retrofit adapters between Smeemra and Hermes. Yeah, there's there's customer base uh cust and customers who are excited about using this data and it's coming out the machine. Yeah, because it in a sense, even though it's line to line uh machine to machine, sorry, yeah, uh it's the first API that you can get from a machine. Yeah, yeah.

Data For AI: From Hype To Use

SPEAKER_01 4:22

So a couple of things to take away from that. I absolutely agree that we didn't address the lack of digital talent we had in the industry at the start of the process. The whole industry 4.0 idea was way too cumbersome, way too large for people to say, actually, I want to do this bit, and this is the ROI, this is the digital dividend I will get from that. So that kind of delayed us and slowed us down. Fast forward to where we are now, where we are the last few years. We have suddenly seen an influx of Stanford graduates into the industry, drawn in by the AI, the data from the ground. Yeah, and we're now in a position where perhaps all that data we do have bubbling up to the surface is exactly what is enabling the you know the potential adoption of AI. And you know, perhaps that's why there's a uh a rush to get that last bit of connectivity, you know, the the kind of SMEMA, Hermes, adapters, those kind of things to get that over the line because we're now saying, okay, well, perhaps we can see our way to a path where we can increase efficiency, reduce um, reduce waste, um, you know, raise the quality bar, lower cost, but by using AI on top of the layer of data that we've produced. And we definitely weren't really talking about we were talking about the self-healing line, but we really, it wasn't an AI discussion 10 years ago. Do you think that's how things have changed? And you do you think we're now on a cusp of a kind of a new, more accelerating phase of developing?

SPEAKER_00 6:04

Of course, yeah. This is a snowball effect, this is a hockey sticker effect. You know, we we it it it was not going to happen in four years or so. Let's say it was it was definitely gonna take 15. But um as again going back to people, there's a generational change too that's occurring. And so I hope so. No, I really do hope so. So tech so the technology is sort of aligning with the people, and so uh you know, there's there's also you know, there's lots of solutions out there that are needing to be re-reinvented or re-created in a sense, uh, and so when that occurs, then it's the communication between the people of one side of the fence and another side of the fence. Yeah, so I remember selling websites when I first got into tech, yeah, and it was going to a golf club and saying, You need a website, yeah. You know, you need to show the customers when the golf club was open and all that stuff, yeah, and book their tea times and do all that kind of stuff. No idea. But yeah, yeah, the the people the people running the golf club, no, didn't know what we've got a big book that does that. And now suddenly we've got forwarding on 20 years, we've now got people on both sides of the fence who understand the common terminologies and technologies and things like that, so it's it's far more quicker to implement things just by simply communicating to other humans, yeah. But certainly we need to sort of bring us all together, yeah, and rebuild some momentum.

SPEAKER_01 7:23

And I I I I really do see AI as doing that, and I'm curious when you say there's a lot of interest in what you're doing, particularly with those interfaces, and people are excited to talk to you about stuff when they're talking to you about that digital transformation, which ostensibly it is, yeah. Are they starting that conversation with AI? Are they starting that conversation with we've realized we've got to use AI, we see some potential, we don't quite know where to start, but we know we have to be connected. Let's get that done.

Stepping Stones To The Connected Factory

SPEAKER_00 7:55

See, I think you're you're approaching that from a management point of view. Yeah, because the people I talk to are the engineers or the production engineers, and the excitement that I'm seeing is people just getting things done. Yeah, you know, being able to move data from one point to another and then trigger something else, or have a notification. And so it's this kind of step-by-step approach. Yeah, but I thought uh Hermes, CFX, both of them, is how you're gonna see it. Yeah, rolled out gradually. The takeaway I've taken from the show today is we need to now seek uh customer success stories or success stories in general of these technologies, yeah, because actually it's going hidden. Yeah, um, I I've not seen many logos of CFX or Hermes on the trade show this year.

SPEAKER_01 8:39

Not many industry 4.0 logos replaced with AI logos.

SPEAKER_00 8:43

Yeah, exactly. You know, it's a it's a it's a master, best marketing. Yeah, a master. So so um so we need to keep on the kind of marketing kind of energy of the communication, yeah, absolutely.

SPEAKER_01 8:56

And I think you're right, when I look at when I look at that AI journey, I think a lot of it does start with solving the problems the engineers want solving, and sometimes that's as simple as connecting two adjacent machines and using what I call adjacent AI to have the SPI maybe make adjustments on its own with the screen printer and that little partnership to work really, really well, and that requires obviously, obviously, uh a seamless transition of real-time data in between those. Um, but then I think you know there are stages beyond that where that gets trustworthy, it becomes autonomous, then that gets pushed up further, and then you have this idea of persona-based agentic AI that sits on top of that, and then you know, obviously that's rising up into the C-suite and figuring out how it can actually impact the business, impact the top line by maybe being able to get my quotes out faster, maybe being able to target customers I couldn't quote um in the past, and then trickling all the way down to efficiency and driving driving growth in the bottom line. And I think because we didn't know what we wanted to do at the start, and we couldn't quantify the ROI, it was very difficult for people to make the I, the investment.

Collaboration, Momentum, And What’s Next

SPEAKER_00 10:09

So yeah, and it still is from the uh equipment manufacturers' point of view. Yeah, uh, some the the issue we're we're we we find ourselves in sometimes is uh a lot of this simple connectivity stuff has been done years ago. We've now what we could consider bespoke solutions because they're not standards, yeah. So they've actually gone you know into a into a they've used a platform, but it is a bespoke platform where now uh some like Hermes, for example, is what I would call a stepping stone standard, yeah, and then the CFX is the connection to the bigger factory, yeah. So it's uh machine builders are also finding this challenge at the moment of of justifying yeah this step.

SPEAKER_01 10:51

Yeah, it's fascinating. For me, it's all about communication and collaboration. I think people like you and I have just got to keep shouting from the rooftops and getting as much information out there. Organizations like uh Global Electronics Association play their part by bringing people together and creating collaborations and serving the industry, yeah, yeah, taking the pulse of the industry, getting everybody talking together, and and as I say, getting data on people out of silos and working together. Thanks so much for stopping by.

SPEAKER_00 11:18

Thanks for all the work you're doing, and we'll um we'll talk again soon.