On this episode of Manufacturing Matters, Ed Goffin, VP Product Marketing at Pleora Technologies joined Jimmy Carroll at Automate 2023 to discuss adoption of artificial intelligence (AI) technologies in the manufacturing space and Pleora’s approach to AI. Specifically, Goffin discusses the use of AI as a complementary technology to traditional machine vision, not as a replacement, and goes into some application examples, including those where the technologies could help a human operator make a decision consistently, repeatably, and in a traceable way.

He also discusses ease-of-use in AI and lowering the barrier of entry, successful real-world AI deployments from Pleora customers, the growing importance of industrial automation technologies in the world today, the latest and greatest advancements and technologies seen at Automate 2023, and more.

Jimmy Carroll: [00:00:06] Hi everybody. My name is Jimmy Carroll. I’m the vice president of operations at Tech B2B Marketing. I’m here at day one of Automate 2023 in Detroit, and I’m joined by Ed Goffin of Pleora Technologies. Ed, thanks so much for taking the time.

Ed Goffin: [00:00:17] Yeah, thanks.

Jimmy Carroll: [00:00:18] So for those of us out there in the world who don’t know about Pleora, tell us about what the company does and what you do there.

Ed Goffin: [00:00:26] Yeah. So I’ll start with what I do first. I’m the manager of marketing, which in our world goes from everything from marketing the products to dealing a lot with customers. As a company, we’ve been in business for 23 years. Probably a lot of people would know us as a component supplier. So we provide frame grabbers, embedded interfaces, hardware, and software that provide a connectivity solution for an imaging device, whether that be a camera or an X-ray panel to processing. And over the last couple of years, we’ve taken some of that real-time imaging expertise, processing knowledge, connectivity and started to move more into the manufacturing space with a couple of solutions for inline applications and offline applications.

Jimmy Carroll: [00:01:11] I followed the company for a long time. I visited you years ago, and over the past couple of years, a big focus for the company has been AI, and so much so that you told me here at Automate that’s a major focus for the company. Can you talk to me about that and what exactly it means for the company and what you’re doing there?

Ed Goffin: [00:01:28] Really, when it comes to AI and the manufacturing world, we’re really concentrating on two areas. One is around applications where there’s still a human involved in the process and that manufacturer doesn’t want to replace the human. They just want to use technology, usually a combination of machine vision and AI, to help that operator make the right decision consistently, repeatedly, and in a traceable way. And then the other application area is in inline applications, where we have a development platform that lets you mix and merge machine vision, AI, open source tools and bring it into a deployment environment.

Jimmy Carroll: [00:02:06] Let’s talk ease of use. So is what you’re doing there, is it designed to make the adoption of AI easier for people?

Ed Goffin: [00:02:13] Yeah, in both areas. So when it comes to a manual inspection application, it’s an app-driven platform. It looks like what you expect from your mobile phone when it comes to apps, right? Really simple to use, simple to program. It’s a drag and drop platform. If you need to customize an app, build your own app. And when it comes to the operator, what’s a little bit unique about our platform is the AI model training is really operator driven. So as they’re making a choice of, yes, that’s a defect; no, that’s not a defect, behind the scenes, transparently, they’re training an AI model that then within a few inspections starts to make suggestions for them. AI thinks this is a defect. When it comes to the inline space, again it’s around ease of use. So simplifying the development of these algorithms and then simplifying the deployment and the integration with cameras, processing, PLCs, MES-ERP systems.

Jimmy Carroll: [00:03:17] I think it was maybe seven-ish years ago, and maybe it was around the Vision Show in Stuttgart, where AI and maybe more specifically deep learning started to be hyped as this technology that’s going to come out and change the machine vision world. And it really hasn’t transpired that way. However, in the past handful of years, there is a set of applications in the real world where deep learning does add real value. And I think a lot of that is working alongside these so-called traditional, discrete, rules-based machine vision algorithms. What are some ways that you’ve seen people have the technologies complement each other to solve their problems today?

Ed Goffin: [00:03:58] Yeah, when you’re talking about hype and AI, it’s interesting. Because a lot of people come to us on the manufacturing side and say, I want AI. And then their next question is: What is AI? So we spend a lot of time educating and training around AI and the difference between AI and machine vision. In the end, they just want a solution that works, right? They’ve got a problem. They’ve got a defect. They’re agnostic to whether it’s machine vision or AI. Where we usually get involved, there’s a bit of a hype cycle of AI, right? They’re excited about AI. They see some perfect use cases. They’re like, okay, that’s my perfect use case. I can do this. And then quickly start to run into problems around: It’s a little bit complex. It can be a little bit costly, especially if I’m working with an already installed infrastructure. And then third and sometimes overlooked but really important is your employees too, right? You need your employees to adopt this technology. As humans sometimes we’re reluctant to adopt new technology. I work in technology. It’ll take me forever to change my phone. I love my phone.

Jimmy Carroll: [00:05:11] So you’re not a big TikTok user?

Ed Goffin: [00:05:12] I’m not a big TikTok user, right? And companies sometimes forget, like, I need the human to buy into this process too. Part of the approach we’re taking is, yes, I understand you want to get to a full-scale automation at the end of this, but maybe there’s some scalable ways you can get there. So instead of jumping to a full automation solution, let’s look at digitizing some of these processes, you know, adding decision support and then gathering that data. And then with that data that can then help guide: What am I going to do next for automation? It also gives you the data to help do things like train an algorithm. That’s huge. Well, every time you start with a new customer and you’re like, I need so many good, so many bad. They’re like, well, I don’t have that database. If you digitize your process first, you can start to build a library of those images and defects. Then you need to be able to move into an inline application.

Jimmy Carroll: [00:06:08] You mentioned perfect use cases, and I won’t put you on the spot to say what’s a perfect use case, but what are some ways that your customers have deployed AI successfully and looked at it and said, “This really improved our operations”?

Ed Goffin: [00:06:19] Yeah. A lot of applications around electronics assembly, where in our case it’s a lower-volume, higher-value type of product, so for them to fully automate that process would be too expensive. But there are high-value parts. They use a human operator to now visually inspect the part. Instead they can use a camera-based system with a mix of machine vision and AI that starts identifying those errors for the operator. They don’t want to remove the operator. There’s a reason that they’re keeping the human in that process, but they just want to give them the tools to be able to see those defects consistently over a long day. We’re doing work with an electronics company. We’ve done some work with a distillery where, again, they have a labeling issue. They have a mix of automation and manual processes in their packaging. They’re finding if something goes wrong with an automation machine and the packaging, they’re not noticing it until it’s at that final packaging stage, and then their CTO is explaining. Basically production stops. All their employees are gathered around looking at this bottle, trying to decide, is it in spec, out of spec. And he looks and he’s like, “My God. Production has stopped. And let’s get it back up and running.” So it’s things like that where he can avoid that downtime. And take away some of the stressful decisions for his employees as well.

Jimmy Carroll: [00:07:43] Not just in your world, your company, but in the overall automation space, what are some ways that the technologies have changed, maybe become more approachable even in the last year or two?

Ed Goffin: [00:07:54] Yeah, I think there’s a huge emphasis on usability, which is a big difference when I think back 10 years ago in machine vision. It could be complicated to integrate these parts. You know, there’s a lot of black magic that happened even around the cameras themselves. There’s way more emphasis on making things usable for people. We’re only at the end of day one. I don’t know how many times I’ve fielded that question of like, “Okay, can I do this? Is this simple enough for me or my operator to be able to do this?”

Jimmy Carroll: [00:08:31] What else are you really excited about here at the show? Again, not just for your company but for machine vision and industrial automation in general?

Ed Goffin: [00:08:38] Yeah. I mean, for one, last year the show, it was a great success, but it still felt a little, for lack of a better word, COVID-y a little bit. I feel like we’re maybe past that. So it’s the networking. There’s way more people. I mean, the show has expanded from last year, and last year it was a big show. Lots of traffic, lots more international traffic. For me, I find it interesting to see, it’s sort of these worlds that I think a few years ago were really separate. The robotics world and some of the tools around palletization and economies of scale for users in my workforce were all sort of separate. And that machine vision, we were sort of our own island out there, and it’s all sort of coming together. So it’s a show like this where you can see a robot in action and be like, “Oh, okay, you know, I recognize a vendor camera on there. I can see what they’re doing. I can see some vision processing happening over here. And this robot is doing whatever it’s doing.”

Jimmy Carroll: [00:09:39] Yeah, it’s really nice to go around and walk on the show floor and see a lot of these vendor companies working together. So components working together to solve unique problems. And you mentioned COVID, and the last few years have really highlighted the importance of industrial automation technologies when we have COVID 19-related disruptions and the great resignation and these continuing labor force issues that for some it’s causing their businesses to close, and industrial automation can in many instances help prevent that. You know, I imagine that either you or somebody from your company was probably at the A3 Business Forum earlier this year. And Alan Beaulieu, during his global economic outlook, said that in times of economic uncertainty, automation will help keep the economy afloat. And I think that what you’re talking about here, looking on the show floor, is a great example, like this is the place to see exactly what people are talking about. From whether it’s a logistics and warehousing application to a more straightforward machine vision application to some of these walking robots that are moving down the aisle to do to different tasks for energy or oil and gas or reading meters and things like that. What I’m getting at is there’s just a lot of ways that people can use automation technologies to solve problems for them and help keep the lights on and boost productivity and drive revenue. And like you said, it’s very exciting.

Ed Goffin: [00:11:04] Yeah. I was at a show a couple of weeks ago in western Canada, and it was primarily focused on natural resources, and their interest in vision was amazing. I mean, they’re coming from an industry they’ve never really adopted a lot of vision in that. It’s a very manual, sort of machine-driven business. And their interest just in vision was phenomenal. Like we had two days of just talking about machine vision to these people, trying to figure out how to adopt it. For those reasons: They’re having trouble attracting operators or employees. They’re looking for economies of scale. They’re trying to reduce costs.

Jimmy Carroll: [00:11:41] That’s a good point. It’s a really good point. I mean, yes, a lot of what we’re seeing here is designed for the factory floor, but these technologies have gone far beyond the factory floor. I mean, if you look at A3 market figures, growth areas are in the service industry and in agriculture, in places like that. So it’s very cool to see these technologies adding value in different ways and including in the field or in the sky, you know, multispectral cameras that may be looking at a field to assess the health of the crops or whatever, or a vision guided robot that picks apples in a field. There’s a lot of cool stuff like that.

Ed Goffin: [00:12:20] A couple of years ago, I always found it hard to, when my kids were younger, to explain what is machine vision. It’s amazing how many more applications are out there in the real world where, you know, now my kids don’t care what I do anymore, but with friends, right, it’s like, “That’s what we do.” Those are the types of applications that vision is helping that company succeed.

Jimmy Carroll: [00:12:41] For sure, it’s not just inspecting a bottle on the line or reading a barcode, which I’m not downplaying the importance of those applications by any means. But there’s just a lot of new applications as these technologies that companies like yours advance, as all these technologies collectively advance, it opens the doors to new applications, and I get to see them on the show floor and read about them and write about them, and it’s very cool.

Ed Goffin: [00:13:08] Yeah, we get to go sometimes on the manufacturing floor and see them too. It’s really cool to see a solution adopted. It’s really cool to learn the problems that they’re trying to solve and looking at technologies. You know, it’s a fun business to be a part of.

Jimmy Carroll: [00:13:23] Absolutely. And thanks so much for taking the time. If people want to learn more about what you’ve been doing lately, what should they do? Should they visit your website?

Ed Goffin: [00:13:31] Go to pleora.com and you’ll see, we’re still investing in the connectivity business. There’s a lot of exciting things going on there for the more traditional machine vision space. And then you can see what we’re working on in the manufacturing side as well.

Jimmy Carroll: [00:13:47] Very cool. Again, thanks so much for taking the time away from your booth on the show floor. This has been a great talk. And thanks everyone for watching the Manufacturing Matters podcast. If you have any questions, comments, concerns, or if you’d like to join:  www.manufacturing-matters.com, Thanks very much.