Episode #71: Fighting Counterfeits: Alitheon CEO Roei Ganzarski on Authenticating Parts
In this episode of U.S. Manufacturing Today (powered by Veryable), host Matt Horine interviews Roei Ganzarski, CEO of Alitheon, about the trillion-dollar global problem of counterfeiting and gray-market fraud and why it is a safety and national security issue in industries like aerospace, automotive, and healthcare. Ganzarski shares examples including end-of-life aircraft engine parts sold as new and fake titanium used in aircraft, and explains Alitheon’s FeaturePrint technology, which authenticates individual items using standard cameras by capturing unique, random manufacturing “tolerance” features to create a digital fingerprint with zero false positives.
Links
- Roei Ganzarski on LinkedIn
- Alitheon Website
- Alitheon on LinkedIn
- Matt Horine on LinkedIn
- Veryable Is Revitalizing U.S. Manufacturing
- Sign Up on the Veryable Platform
Timestamps
- 00:00 Podcast Intro
- 00:41 Counterfeit Crisis
- 01:44 Fraud and Gray Market
- 04:21 Career Path to Alitheon
- 09:11 Aerospace Scandals
- 11:44 FeaturePrint Explained
- 18:15 Why Labels Fail
- 23:48 Wear and Tear Handling
- 28:03 New Manufacturer Promises
- 28:59 Camera Based Traceability
- 29:17 Warranty Return Fraud
- 32:15 Fingerprinting Against Fakes
- 35:29 Supply Chain Risk Trends
- 38:13 Supplier Competitive Edge
- 40:04 EV Battery Safety Risks
- 43:10 Compliance Beyond QR Codes
- 44:30 Blockchain Physical Gap
- 47:28 Proxy On Proxy Failures
- 51:10 30 Day Action Plan
- 52:56 Pilot And Production Setup
- 55:19 Traceability For QA Recalls
- 57:44 Podcast Outro
Episode Transcript
Matt Horine: [00:00:00] Welcome back to U.S. Manufacturing Today, the podcast powered by Veryable, where we talk with the leaders, innovators, and change makers shaping the future of American industry, along with providing regular updates on the state of manufacturing, the changing policy landscape, and more.
Counterfeiting isn't just a knockoff handbag problem anymore.
Estimates put the global counterfeit and gray market problem in the trillions of dollars, and in manufacturing, it's not just a brand issue, it's a safety issue. Fake components have found their way into aircraft, vehicles, and industrial equipment, and with a heightened sense [00:01:00] of national security, this is an, a very timely topic.
And my guest today is Roei Ganzarski, CEO of Alitheon, a company that solved the authentication problem in a genuinely novel way. Roei, welcome to the show.
Roei: Thank you very much. Happy to be here
Matt: We are very excited to have you because this seems to be a problem that comes up in a lot of different ways, and you and your company have found a way to address it with an incredible technology.
We got to know each other not too long ago and was very impressed by it. But before we get into the tech and the company and the real problem and how big it really is, where does it show up worst in manufacturing and how does it show up? Do those numbers sound right to you?
Roei: Yeah, no, a-absolutely they sound right, unfortunately. And it's interesting, when people think about counterfeiting two things usually come to mind. One is really the fake aspect, someone made an item to look like mine for various nefarious reasons, financial, national security, et cetera. But two, people automatically think, the fake [00:02:00] Rolex watch and the fake Louis Vuitton bag that you buy for 50 bucks on Canal Street in New
York or in Thailand or something like that. And that really isn't the problem, and it isn't the only problem. Really, if you think about it we look at the world as fraudulent activity. Fraudulent could be counterfeits, fakes. It could also be gray market. For example, real but expired pharmaceuticals being sold with a label that says they're not expired. These are real items. They're authentic. There's nothing fake about them except the date has been manipulated. So regular anti-counterfeiting solutions won't work here 'cause these aren't counterfeits. So that's kind of one aspect. The second is the topic itself. People think about, again, the watch, the bag. Those are in fact a problem, and it happens all the time. But the bigger problem is where health and safety come into play. Fake brake pads that are sold to automobile owners used end-of-life aircraft engine parts being sold to airlines [00:03:00] with paperwork that say they're brand new, things like that.
And those problems, when you think about it, lead me to the conclusion that greed has no boundaries. If you're gonna, you're gonna cheat someone out of their money by, by providing a fake bag that's wrong, it's illegal, that's not okay, but there's no physical element to it in terms of risk, safety. When you're introducing fake medicine, fake brake pads, stolen stuff, you know someone's gonna get hurt,
and still you do it. That, to me, is crazy, and that's really the mission that we're after a-at, as Alitheon. And yes, that problem is big, it's real, and people might not like to hear it, but it's right here in the US, it's in Western Europe.
This isn't a Africa, China problem. It's a global problem, and it hits us in the US and Europe just like it does everywhere else.[00:04:00]
Matt: Boy, you've really probably touched on a couple of things there. The safety aspect of it, the the ability to-- There's so much in the world today that really is, there's no other way to describe it, it's fake. And it's for greed, it's for profit, it's, for whatever purpose. The, there was an article that came out this week that a lot of things that were marked through the GSA were, made in America, actually made in China.
They made their way into it. That was new since we met, but something I certainly was, doubled down on, wow, this is an extremely timely topic. But before we get into the technology and how all of it works, you've had a a great and genuinely interesting career arc into, some of the, some of these insights that have framed your point of view and what you're doing now.
The over a decade at Boeing leading the charge in, other avionics and, or aviation spaces now Lithion. Walk me through that path a little bit, because I think it really frames where you're coming from on these fake parts, 'cause you did highlight the airline industry, and that's top of mind.
Roei: Yeah, I have been blessed and lucky to have been given a lot of interesting opportunities and then thanks to my parents saw that there was an opportunity and grabbed it as well, right? I like to talk to students a [00:05:00] lot here at the University of Washington and other places, and I tell them, "We're all lucky at some point in time," right?
What differentiates people is one, can you identify that you're being lucky and two, then take that luck, right? 'Cause we're all gonna be
lucky at some point. So yes, I've been blessed and lucky i-in these opportunities. So as you mentioned a long time at Boeing specifically on the services side, so everything from training of pilots and mechanics to software to spare parts, et cetera. Then I left Boeing when my wife claimed I was having a midlife crisis at the age of 40. Left Boeing and went to the world of startups. First one was around the use of mathematics and heuristics to do real time optimization of on-demand schedules. Then it was back to aviation with electric aircraft both electric propulsion systems or engines for aircraft and electric aircraft themselves, and now with Alitheon.
And really w- I like to think of it as every step in the way I added something to my knowledge, my learning, both in [00:06:00] what I wanted to do, but also what I didn't wanna do. What I wanted to see in a company, what I didn't wanna see in a company. Specifically in the realm of where I am today with counterfeiting fraudulent goods, gray market, et cetera, if you think of the path through aerospace hardware highly critical items two kind of stories come to mind.
One is when I was in the electric propulsion company, we would get parts from suppliers, and they would come with reams of paperwork that said, "Here's what the part is, here's how it was made, here's how we tested it, here's what we did to FAA compliance," et cetera, et cetera. And I remember some points, even before knowing about this problem as a problem, I would ask the supplier, "Hey, how do I know this ream of paper that's all great data belongs to this physical part you just gave me?"
And the answer was, "Oh, look, there's a tag that's tied with a zip tie to the part and you can see that it belongs to this thing," or, "There's a sticker or a serial number etched on the part." And I would naively ask, "Couldn't anyone etch that serial number or couldn't that tag fall off?" [00:07:00] And the answer was, "Oh yeah, but no, th-th-that probably won't happen." But it does happen, a lot, intentionally and unintentionally. So that was one thing that was in my mind for quite some time. Another is for w- our 25th anniversary, my wife bought me a watch that I really w- was my, one of my dream watches. Beautiful watch made by IWC Portugieser.
G- gorgeous watch, and my wife surprised me and bought it for our anniversary. And at some point afterwards, I had to go and get it serviced, right? You give it for service, maintenance, et cetera. And I went to the store and said, "Hey, I'd like to have this watch serviced." And they said, "Oh, great, we'll take it, and come back in six weeks and you'll have it back." And I said, "What do you mean in six weeks? I thought you were gonna do it like now, like I bring my car for an oil change." "Oh no, we have to send it to the service center," et cetera. And I naively asked, "How do I know I'm gonna get it back?" This is, not only for me at least it was an expensive watch, but two, it's a sentimental one.
How do I know I'm getting it back? And the guy in the store said, "Oh, just look, there's a serial number." Now, we've all been, to our earlier comment, we've all been to the Canal Street, to Thailand, to [00:08:00] China, where you can see these watches replicated one to one, including the serial number. And I was "How do I know?"
He said, "Oh, you could take a picture of it," but that wouldn't help me. So those two things, both, both on a professional and personal basis, were on my mind, and when I got to see the Alitheon technology that was being developed in its early days, to me it was obvious, wow, this thing solves both problems. The tracing of a aircraft part and knowing that part, in fact and indeed, is correct and belongs to the data you have. But two, on a personal note, oh, I could know that the watch I gave to service is the watch I'm getting back, not one that looks like it, that parts have been changed, et cetera. And so for me, that was the trigger that said, oh, wow I have to go on with this Alitheon thing and make it happen, and that's what I've been doing for the last four and a half years.
Matt: Yeah, those are some really concrete examples because you do, in your daily life you think about all the times where you could interact with something like that not getting something back that is [00:09:00] personal and sentimental. And then from the aviation standpoint specifically, it's highly concerning, especially some of the, the work with, the elec- electric aviation and those types of things.
It gets much more technically and technically complex.
Roei: At just the past few years, in aerospace alone, in the past few years, two really kind of global cases came to light, and these are just the ones that were caught, right? One
was a legitimate UK supplier, again, not a supplier that people think fakes come from China or Africa. In the UK was selling end-of-life used aircraft engine parts with paperwork that said that they were brand new back to the airlines. And airlines like Delta, Virgin, United, Southwest, were all I wanna say caught, but found flying with these parts in their engines solely by the grace of God as, as I'll say, that this wasn't caught by a hull loss or an accident with multiple people dying. It
was caught by a mechanic who, by mistake, thought something was wrong here and found it after over four years of [00:10:00] this company selling these parts out to the marketplace. Unbelievable. Fake or not fake paperwork saying this used end-of-life part is good. Now, someone might say, "Hey, what's the problem? I buy a used T-shirt, a used watch, a used..." Yeah, but engine parts have a life to them because of stress tensile strength that, it won't stand up to the stresses of being in an engine after a certain amount of time, so you have to take it off the engine and destRoei it so that no one uses it.
These guys just cleaned it up, put up paperwork that said it was real, and sold it back to the very airlines they'd gotten it from. Just amazing. The second one, Boeing and Airbus both were found installing fake titanium on their aircraft that they got from some, again, illegitimate supplier. So the paperwork said this was good titanium, but the titanium itself wasn't, and it was installed on aircraft. These are just two that by chance were found. Who knows what isn't found? And this is just one industry, one example,
and you can, in every industry, find counterfeit [00:11:00] HIV pills sold in the United States, counterfeit Ozempic syringes sold in Europe, all sorts of examples of things like that are just detrimental to health, let alone, of course, the fake watches and bags that you hear about coming into the US and Europe all the time. That of course happens, and some of it gets caught, but the bigger ones is just on a daily basis
Matt: It's an incredible real world examples because it's one thing for, getting ripped off for consumer product, which is like a, it's a distinct category of its own. But when you get into safety and zero fail environments like like aviation it's pretty scary. It's making me double down on my John Madden theory about just driving across the country.
I'm trying not to, I'm trying not to take that route, but this may be feeding into it a little bit. Good segue into, getting concrete on the technology and what your company is doing. FeaturePrint, it authenticates a physical part using just a standard camera, no tags, chips, labels, QR codes.
QR codes are something that everybody's become very familiar with. Tell us a [00:12:00] little bit about how that actually works, because when you described it to me, it was almost as simple as taking a picture. But would love to get some technical detail.
Roei: So first of all, it is in fact as simple as taking a picture. The premise behind it or the underlying kind of foundation behind it, and since this is a manufacturing-related cast, everyone will understand this, machines cannot do the same thing over and over again nominally. That's why,
as mechanical engineers, manufacturing engineers, design engineers, you give the machine or the system tolerances that will enable the quality control to accept or not accept a part.
So for example you make a brake pad, an engine part, a watch, it doesn't really-- it doesn't matter what. Everything that's manufactured will get a tolerance. It might be I need this item to be or this scale to be two millimeters plus minus three-thousandths of a millimeter. Why? Because as long as the machine manufactures it within that tolerance, within that kind of Six Sigma quality control band, the part passes QA, it [00:13:00] functions as it should, it looks identical to everything else, and it goes out to the marketplace, right?
If it's out of tolerance, then you fail it on QA, it goes back to rework or whatever it is. But as long as it's within that tolerance, it's good. So it's not nominal. It's never gonna be nominal. There's always gonna be tolerance aspects to it. So it's not a quality problem with... as long as it's within the tolerance.
It's just a manufacturing aspect. What we have developed is machine vision algorithms, basically mathematical and physics-based algorithms that, using a camera, can take a picture of millions of identical items, literally millions of identical items, and see, quote unquote, or find all those little elements that are within the tolerance band, but not nominal And they are by definition chaotic and random, right?
The machine doesn't design this point to be plus one hundred and twentieth, this point to be minus one hundred and thirty-second. It's just random and chaotic by the nature of manufacturing. We have developed algorithms to [00:14:00] see it, codify it into a digital fingerprint of sort, we call it a feature print, and then store it. And then in the future, someone takes a phone or a camera, takes another picture of that item, another digital fingerprint or feature print is created using the same machine vision algorithms that we developed. And then we do a search. And if you watch CSI or Bourne Identity and they're looking for a,
a human fingerprint, they're comparing not the images, but the data from the original fingerprint and the now found fingerprint.
We do the same thing. We'll load that newly created digital fingerprint or feature print to our servers in AWS they sit, and we'll do a search. And we'll either find that item, not an item like it, not the class of item, but that item, or we won't. So we'll tell you hey-- your listeners can't see it, but I'm holding my cup of coffee here. I can identify this cup of coffee as Roei's cup of coffee, the cup number one seven six dash four, [00:15:00] not dash three or dash five, but dash four. Or, "Hey, I've never seen this cup of coffee before in my life." So that's kind of the premise of what it does. So if you think about that The likelihood, and we find, by the way, thousands of these points. The likelihood of having two items, even off the same production line, done one after the other, have the same set of in nominal or non-nominal tolerances, again, not a quality problem, still within the band, but non-nominal, having them be the same between two items, the likelihood of that is one to six plus trillion, which makes a feature print, our digital fingerprint, as unique, if not more than a human fingerprint. So from that perspective, it's mathematically proven. The key of what we developed isn't that we use this idea. Many have tried this idea of the fact that there's nominal manufacturing bands is not new, right? Anyone in manufacturing know this. What was unique or what's key about our technology or our algorithms is, one, we [00:16:00] do it with standard off-the-shelf cameras. No spectral imaging, no infrared light, no special invisible tags or dusts or markings. Literally, we take a picture of the item with an industrial camera or even a mobile phone, and that's it. Within-- Once that ca-- picture is taken, the digital fingerprint is done, and you can immediately identify it, even at the continuation of that production line, for example, for in-line tracing. That's one, traditional standard off-the-shelf cameras. Two, I don't need to know what the item is. In fact, with most of our customers, we don't know what they're feature printing. They, they just do it in their line. So I don't need to know what the nominal and tolerance band is. Our algorithms know to look for the, quote-unquote, "chaotic, non-normal nature of things" and are able to categorize it that way.
So from that perspective, that's good. And then three, we're able to, in production, get to zero false positives, which is really critical. And that was a rule of ours, if you will, when we developed this mostly coming from healthcare or safety-oriented industries. Let's just take [00:17:00] healthcare as an example.
If I tell you this medicine is fake and I'm wrong, meaning I should have identified it as one of yours, but I'm telling you, "Hey, I don't know what it is," that's a false negative. You lost a couple of bucks, right? You throw it away or check it maybe another way. But if I tell you fake medicine is real and I'm wrong, false positive, that could be detrimental.
And so we said we have to be able to, in production, get the false to zero false positives. So our algorithms, our system, is tunable to the actual products or things that the customers are using. So those are key elements to it. It took us about four and a half, almost five years to develop this key kind of deep tech capability, and today we go to market with it in multiple industries.
It's actually pretty cool to see.
Matt: Yeah. No it's a great thing to, one, the vision of it, and two the practical application of it because it is so unique and you're doing it from, the surface of an object. This isn't something that's, dissected. It's finished product. It's component parts. It's all of those things.
It takes me back a little bit more [00:18:00] to time on the shop floor or being in procurement or ordering parts. I've have my list of favorite vendors and my list of least favorite vendors, and no names will be mentioned, and it's all for one reason or another. Generally pricing, kind of those types of things for me.
But also, the quality aspect of it. Is it safe to say that removing the tags and labels and all that other, I don't wanna call it junk because it, it plays a core part in tracking as, the current state, QR codes, is removing all that making it more secure, not less?
Roei: It depends how you look at it. So one, I think there will always be a need and a place for some sort of proxy label.
And the need and place will be conveying information. So let's, take the parallel, which is some of the inspiration is human fingerprinting or biometrics, right?
People's biometrics. Today, you travel internationally or you wanna go into some secret room in the Pentagon you give your passport, right? But then what do they do? They take your passport, but they also take biometrics, fingerprints, facial, [00:19:00] iris, maybe all of them. Why? Because the passport is a really easy, clear way to convey information.
"Oh, I see your passport. You're from the US. Here's your date of birth, color of eyes, where you were born," et cetera. But it w-- but when they wanna know, does that passport really belong to you, they take your biometrics. Why? Because governments have figured out, as security-related organizations, passports, tags, badges, IDs can be manipulated, changed, even faked. But your biometric is fairly robust to identify you as a person. Imagine an extreme scenario identical twins using each other's IDs to get in somewhere.
The individual is real, the ID is real, but the ID to int- to physical individual link is not. If you only rely on the proxy, you might look at the person, look at the proxy and say, "Yeah, you're good to go.
Come and enter or travel," et cetera. But if you take their fingerprint, identical twins or triplets have different fingerprints, even though they are identical [00:20:00] in view, then I could say, "Hey, this fingerprint isn't what's represented in this ID. You are not the person this ID represents." So there's always room for label, but with people we've moved to biometrics. We're simply doing simply, quote-unquote, biometrics for things. We're basically saying the label says, this is a brake pad made by this company a-and sent to this car on this date, or this healthcare product is made by this company, it has an expiration date of this date, et cetera, et cetera.
This is the distributor it's going to, and so on. That's what the label could say very easily and very simply, so every human being can look at it. But our feature print will then say, when you look at it and you know to, to do it, the label actually is telling the truth about this product. We, for example, some of our customers have found i-in healthcare, believe it or not, their own products being sold... I won't say, give the exact example, but let's say, expiration dates may be modified diversions within markets [00:21:00] modified, authentic products being sold illegally to the detriment or the risk of the end user. But the products were authentic, but maybe they expired. The products were authentic, but maybe they were being sold in the wrong place. And the labels don't help you on that because, hey, I could create a label that just tells something else, and I read it, and the label is correct. It just doesn't belong or isn't correct to that physical item. So labels have their place, but when it comes to authenticity, I think they're not only not good enough, in fact, they can create overconfidence. "Oh, look, this product has a hologram. Oh if it's a
colorful hologram, it must be real because holograms show authenticity." But no, holograms are just a really cool sticker that says at best that hologram is real.
It doesn't mean that what the hologram is on is what it's supposed to be. And there have been multiple cases of [00:22:00] fake products, gray market products, fraudulent products having either real or real-looking holograms, QR codes, barcodes, and so on. So I think when you look at authentication, relying on these proxies are not good. When looking at simply information, they're great. You can in fact reduce your cost. Don't buy these really expensive holograms and things that are supposed to be secure, because the fact that they're not securely attached to an item, having a sticker or an etching means nothing today, especially when you look at the high-end counterfeiters or fraudulent actors in places like China, Russia, and so on. They have the same machinery that the original manufacturer have. They have the same printers. They have AI today to help clone and create perfect copies of a QR code or of a website or of a hologram. But what they can't do is create a perfect copy of the original item, because even the original manufacturer can't do that, going back to the, tolerance bands.
So that can't [00:23:00] be faked. And so from that perspective, if you want true authentication, you wanna do biometrics like you do for people, for things. For information, for, "Hey consumer, here's a nice sticker that makes you feel good," yeah, that's great. I wouldn't rely on that for true authentication, traceability, provenance, and so on.
Does that make sense?
Matt: Totally. That's a great analogy on the passports. And when you talked about the hologram stickers, I always think of, like officially licensed merchandise. If you buy a-- if your team wins a championship it comes with a little sticker on the official hat, and it's kinda that sticker doesn't do anything.
At the end of it, the team is going, if they just won, you're paying probably $50 for a baseball hat that says that they won, right? And then eventually it goes to the discount rack after the time goes on. But the holograms is a sticker. There's nothing that says this was a counterfeit product or anything to that degree.
So that's where my mind went. This may be a little bit more of a, kind of a, somebody new to the technology, so don't take it as too pedestrian a question, but something that came to mind. What happens if a part gets scratched or worn down [00:24:00] or damaged in some kind of way? Or do people un-enroll it from the system?
What are some of the things that you've seen? As, as time goes on, obviously parts change. Fingerprints don't necessarily change, but physical appearance might. So is that something that's accounted for?
Roei: Absolutely, and that's that's actually a really good question and a technical question, in fact if you want it to be. So if you think of human fingerprints same thing could happen, right? You could scratch your fingers. As you grow from child to adult, your fingerprints stretch because your skin is stretching, et cetera.
So there are modifications, and those systems are robust enough to do. You hear them on TV, partial fingerprints, right? Et cetera, et cetera.
Our system finds thousands of points, if you will, that make that item unique, that digital fingerprint unique And it's not because we need all these thousands of points, but it's because we want exactly to be able to address the point you raised. Maybe it gets damaged and some of the, the product itself is now missing or [00:25:00] scratched or buffed up or broken or just gone. We still wanna be able to rely on enough of those, if you will, points, let's call them points of data more importantly, that say, you know what? Even though ten percent, fifty percent, we've seen cases where ninety percent of the face of the item is missing, and our system can still tell you exactly what item that was.
Again, not the class of item, it's a cup of coffee or a computer chip or a brake pad, but it's that brake pad,
that computer chip. And that's really important because in real world, these physical products do get damaged. Again, l- go back to the basic example of a purse or a watch. They get banged around, the purse could get scratched, UV light just from being with it outside. We have done with our customers tests of twenty-five, fifty years of aging and damage, and with a mobile phone, I can still identify products.
We've had customers test it with a grinder, take titanium parts, grind ninety percent of the [00:26:00] face of them, and still be able to identify them. There have been some really crazy luxury clothes going through a washer cycle,
Matt: Yeah
Roei: and still being able to identify them.
The, the thing is really robust. Now, to a point, right? It's not magic,
it's math and physics. At some point, you erase enough of the item, of the face of the item, or you cover it up. It's like I'm gonna sand off my fingerprint or wear gloves. At some point, there is no fingerprint to be seen. So it's not magic here that'll always work, but it's fairly robust to what we've seen our customers define as normal or even extreme, but logical wear and tear.
So from that perspective, it's pretty cool. So that's kinda one part of the answer. But the second part operationally is that our customers can choose to add feature prints, if you will, through time. So imagine, let's go to human fingerprints. Imagine I take your fingerprint when you're ten years old, and then I see you again when you're twelve. I'll take another fingerprint. I won't delete the ten-year-old fingerprint [00:27:00] 'cause why delete it? It's good original data. But I'll take another one when you're twelve, and then I happen to see you when you're fifteen, I'll take another one. Now when you're twenty-five, I'm not necessarily comparing your fingerprint to my baseline when you were ten, I'm only comparing it to the baseline when you were fifteen.
So I can keep adding feature prints. I don't have to delete anything. I can keep
adding feature prints as I see the product over and over again, maybe through service or, you come into the store, or I return... W- whatever it is, I can, as the user, continue to update, if you will- my digital fingerprints, and then the next incremental one is only incremental as opposed to back to the baseline.
So those are two methods that customers can use or see in terms of wear and tear, robustness, and so on. And so far, we've seen it work phenomenally well
Matt: No, those are really good analogies, especially on the actual fingerprinting, right? Your biometrics. They're your, they're data points. It's all kind of relative when you said it's like partial fingerprints, or you think about dental records. I don't mean to be, that, that dark about it, but the dental records are something that's [00:28:00] part of that that change over time significantly.
Really good points on those. I- is, so this gets to the heart of the question, and I think, when we were talking originally, it's about a manufacturer's relationship with their customers and, the general public a- at large, right? And, from the external promise that it makes, what can a company, guarant- or promise that they couldn't before necessarily with this level of of identification?
Roei: Yeah, th-th-that's a really good question, and I'll even expand on that on the premise that it's between the manufacturer and their customers.
We have customers who are manufacturers that use this for inline traceability. They don't even think about going out with it. They're doing it for in-factory, because maybe the items they make, think, high-end electronics and so on sensitive electronics, sensitive products, or products that you simply, you know, like a luxury, a jewelry or watch, you're not gonna start sticking stuff on it or etching it in order to trace it in the factory because either you can't, you're not allowed to, et cetera.
So [00:29:00] we have customers who do it simply for traceability, because all they have to do is take a picture. So you could have, low-cost industrial, we're talking fifteen hundred bucks industrial camera. These are traditional ins-industrial cameras you would hear that companies use, Basler, Allied, et cetera, sitting on a line and tracing things in the factory, let alone outside. What this really allows your the manufacturers to do both in line and when they go out to their customers or their distributors, et cetera, and, by the way, an added use case, when products come back to either be returned or repaired under warranty, you are able to know, is that in fact the product that I made in the line, that went through these specific machines, that I sent to that distributor, that the distributor sent to the customer, that the customer is now returning to me or asking me to repair under warranty?
Is it that product? And I can rely on that from the product information itself as opposed to the proxy Let's say [00:30:00] there's both cases of both unintentional or routine, "Hey, I'm returning something, but by mistake I returned it," or, "I'm asking you to repair it under warranty, and I didn't notice that I was using a product that was already two years old and no longer under your warranty, versus the one I just bought 12-- two months ago and is under warranty." I could be doing it by mistake, or I could intentionally be saying, "Hey, I just bought this." I go out, gra-grab the old broken one from my car that I bought two years ago, put it back in the package, and now come back with the original package and the receipt and say, "Hey, I'd like to return this." You'll look at it, it's your product, and you take it back. But it's not the product you just sold, you've just been scammed, right? Or, "Hey, please repair this under warranty," but that's no longer under warranty. Or someone else tried to fix it, and you can now tell, "Hey, that's not the part I installed. It's a real part, but it's another part," or maybe it's a refurbished part. You can now offer your customers the ability, and yourself, the ability to identify both at the [00:31:00] component and the assembly level, the part level, the product level, simply with a picture that the actual part itself that's either being sold, distributed, returned, repaired, serviced, is what the data says it is. As opposed to, "Hey, I've been using this for quite some time in the field." The sticker fell off. The serial number got scratched. I intentionally moved the marker from one to the other, and now you don't know if the item matches the proxy or the data that you think it does. I'm not even talking about the fact that it may be a fake item.
I'm talking authentic item, but just not the one you think it is. So you can now start doing things at a level of provenance and traceability and authentication and serialization that wasn't possible before. And I say wasn't possible from a perspective of, yeah, m- people might say, "Oh, I do that today with my serial number sticker or an etched laser.
Look, it works." I said, "Yes, you [00:32:00] think it works." You don't know unless you happen to by luck catch it how many of those things aren't what you think they are because the number got scratched, copied, the sticker came off, half of the barcode got erased with a Sharpie and now you can't read it, et cetera, et cetera.
So you can now do it, just like fingerprints, at a level that doesn't expose you. Again, go back to the human being When my teenage kids, maybe, I, of course I'm not saying anything, but when my teenage kids wanted to go to a bar and have a drink before they were 21, maybe according to stories, they had a fake ID, let's say, right? And that fake ID looked really good. There's stories of, fake IDs getting caught or found in Hong Kong on the way to the US of all 50 states that all had the holograms, the anti-counterfeit, everything on them, except the, the card stock wasn't real. But I go to the bar, I show them my fake ID.
He looks at it, it looks real, he said, "Hey, come on and drink." They're still liable if I'm caught serving an underage person a drink, Right
So it [00:33:00] doesn't take away the liability. But imagine, not this would happen for fingerprinting, but imagine in the entry to the bar, there was a fingerprint system, and they say, "Thanks for the ID, but put your fingerprint here."
And then I would know, hey, you're 18, you're not 21, or you're 20, you're not 21. You can't enter the bar. I don't care what your driver's license say, if it's real or fake or whatever it is. So from that perspective, you're now creating an environment as a manufacturer to your distributor, to your customer, to your service agent, to the return, where you don't have to rely on that proxy that may or may not be right because you're still held liable. If a customer gets injured by a product, even if that product is fake, if it has your brand on it, guess who they're gonna blame? Guess who they're gonna come after? Guess who they're gonna go online and say, "That company's products created my injury or created my child's injury"? Now, you could say 100 times, "Oh yeah, but it was fake.
It wasn't mine." Hey, it says your [00:34:00] brand, and you never gave me as the customer the ability to say it wasn't, so this is your fault. That's really what you're now, or what our technology allows a company to do.
Matt: No, that's, it's a, a really good example, and I think too, I've previously worked in an industry where a, a company that made mechanical pipe joints has a very distinct color. Not naming names, but there's a even larger international version of the company that makes similar pipe valves fittings, and they use the same color paint for those couplings, and it was a, it's a branded color.
It actually involved... I think it was quite controversial. There was a, there were some lawsuits if I remember correctly, but it was because they were two very different products doing the same thing, but the competitor had just adopted the same color. And, for most people, this manufacturer made more sense because of the guarantee of the product, the warranty the, the testing of it, the, the quality assurance.
It was a totally real world application that came to mind just because, and they didn't mislabel it or try to [00:35:00] serialize it or do anything. It was just a little bit more of a marketing trick, but you take that down to the counterfeit level and it's something pretty incredible.
For something, for our listeners, we talk a lot about seizing the market where it is right now. There's a lot of big trends in manufacturing. How does this kind of serialization or this type of approach, help competitive advantage? And maybe if you have a customer or someone at top of mind that could, you could give us as an example that used it as basically a selling point of something they were doing
Roei: Yeah, no, re- really good question. So first of all the world is already in and moving more and more to, to highly distributed supply chains highly international supply chains or global supply chains and very highly digital transactions, which means for the most part, you don't always see where your items are made or where they come from, how many hands they've changed et cetera, et cetera. And so where once, you'd walk into a store or into a factory [00:36:00] and you might think, "Ah, you know what? This place looks a little shady. I don't know that I wanna buy here or outsource here or manufacture here," today that doesn't happen. It's really hard. So that creates a phenomenal opportunity for any bad actor to say, "Hey, the likelihood of you finding out is very small, and two, there are so many touchpoints in which I can introduce these gray market fraudulent counterfeit products that, again, the likelihood of you catching them is very small."
And you can see, even with the trillions of dollars a year being identified by International Chamber of Commerce Customs and Border Protection, places like that, of things that they catch, how many things do they not catch, right? It's that iceberg analogy. So that's one.
Two, because the world is moving in that direction very quickly, and the generation coming up now is more aware...
I won't say they care more, but they're more aware of environmental aspects, social aspects. Governments are now regulating more and more various aspects of manufacturing, supply chain, et [00:37:00] cetera. For example, in Europe they're introducing, starting to introduce DPP, the digital product passport. I wanna know for high consequence sensitive items, et cetera, that when you say, "This part was made in this place from these materials recycled," et cetera, that actually you can prove it. Now, unfortunately, they are still thinking about the data as the key element, which is very important. But then tying that data in the blockchain, very highly protected, to the item with a QR code or a barcode. You're opening the door again for bad actors to say, "Oh yeah I'm good with your blockchain data.
I'm just gonna manipulate it through a fake QR code," or moving the barcode somewhere else for example. But it's showing us that regulatory authorities are more and more pushing towards, "I want you to show me what you're claiming." Right? So it's not just, "Oh yeah, this is made in the USA." Show me and prove to me that it's made in the USA, and not by slapping a sticker on it that says, "Made in the USA," and the [00:38:00] sticker, by the way, was printed in China, right?
So we- we're getting more and more to where customers want to know and get proof that this is real, and relying on proxies for that proof is not good. So that's one element to it. The second, from your selling perspective, imagine... I'll just take an an automotive example, right? So imagine you're an automotive brand.
You don't make brake pads, right? You bu- or airbags. You buy them from various suppliers. Imagine I'm a brake pad manufacturer. And I now want to be-- a-and the brakes have all the same specs, right? Some brake pads might be better or worse or better performance, et cetera, but for the most part, let's assume two competitors have the same specs. Brake pads get counterfeited way too much, and very easily there have been multiple cases of both brakes and airbags. So imagine as a supplier, I say to that brand, that auto brand, "Hey, I know you're bidding, a few of us suppliers against each other to get the best deal, et cetera. But if you buy from me, I can also ensure [00:39:00] that you get access to this feature print technology so that you, the auto brand, your service centers, your authorized garages and dealerships, and your consumer can very easily, with their phone, tell that this brake pad is in fact the original brake pad made by our company for your brand for this model car, and not a brake pad that looks like it and has a nice sticker because, hey, bad actors know how to make a brake pad that looks like mine and put a sticker that looks like your brand." So suddenly it's a competitive advantage. Now, whether you wanna charge more for that or not, that-that's the deal of the company. But
as a competitive edge, you're saying not, "I'm gonna give you the same barcode sticker that everyone else does, or the same hologram sticker that everyone else does. I'm going to give you an ability to down the entire supply chain and through the life cycle."
Hey, I removed this airbag and now I'm not just selling it to the secondary market for someone else to think it's new. I can identify [00:40:00] that airbag you just uninstalled or took out as this won't be used. Think EVs today. I guess some of my background is electric aircraft, think electric cars. I once had an electric car, I won't say what brand, and unfortunately I totaled it.
I had an accident. Luckily, no, no one got hurt, but the car was deemed totaled. The batteries Who knows, right? When
that car is totaled, where does it go? Where do those batteries go? Now, coming from electric aviation, I can tell you one bad cell, not a bad battery pack, but one bad cell could increase the likelihood and the chances of a thermal runaway and a fire propagation by multiple percentage elements.
Meaning I am now taking, maybe if I wanna be cheap, I'll take some of these batteries out of this totaled car and I'll install it somewhere else. Not necessarily as the manufacturer, of course, service, secondary upgrades,
updates, et cetera. And then you're increasing the chance of that problem.
Right now, I'm [00:41:00] relying on some printed QR code on the battery pack that says, "This is the battery pack." Lo and behold fake batteries with good, stickers that say or barcodes that say that they're authentic were found in Europe already in the past years.
These things are happening. Anywhere there's demand for something, there's gonna be supply, and supply isn't always authentic. So think of that element of safety, the batteries, the airbags, the brake pads. Now, as a supplier, you can tell the end customer, maybe the end customer is the brand OEM, maybe the end customer is the government, the military, right?
Who knows? But you can tell them, "Hey, if you take my products, I can give you this technology that through the inherent features of the item, not a sticker, a barcode, a QR code, you and your customers and your distributors and your entire life cycle, the, the recycler, can know that's the product." Speaking of recycling, by the way we've seen from many companies that have recycling programs that the various tags, stickers, QR codes that are [00:42:00] added on create problems for recyclers because they stick-- they, they screw up the recycling machines.
So a cardboard box or a packaging that comes in with stickers on them, they throw out because they can't recycle them. So now your recycling program suddenly isn't we're completely recyclable. Imagine if there were no need for stickers, the package itself was what was identifying itself, you could go into recycling really easy.
So there's so many value-added points that as a manufacturer you can offer as both a competitive but as a real world life cycle program as well.
Matt: Yeah, that's, that goes to my theory that a lot of my recycling doesn't actually go to recycling personally. But no, that's, it's a really good point and a distinct competitive advantage. You touched on something there, the kind of the global supply chain, some of that validation, all that. The volatility component of this is so significant that people are trying to stay in compliance with either, Buy America Act or other types of things that go up and down their supply chain.
How much [00:43:00] easier does it make it for, not just quality control, but compliance reporting? We just did an episode a, a couple weeks ago on some of the BABA compliance and some of those types of things. How much does this help that process?
Roei: Yeah, no, i-it's fact- it's not so much helping the process, it's ensuring the process, right?
'Cause I could easily argue, hey, my QR code helps the process of a compliance. Scan the QR code and that's it fills up your compliance sheet. That's great, except I don't know what that QR code is, where it came from, who made it.
Is it real? Is it authentic? Is it actually on the right product? What we offer from a compliance perspective is to know that what you're checking and looking at is what you think you're looking at. That is true compliance, right? And again, just if your compliance program was to make sure that the 10 people coming into your room or into your country are who you think they are based on the information in the passport, compliance might say, "Oh, I just scanned the passport.
Yeah, it shows me your name. Great. Welcome. Come on in." Or, "Here's your [00:44:00] visa. Oh yeah, the visa looks real. Come on in." But is the visa yours, or is the visa the passport's, right? 'Cause I attached the visa to the passport, so yeah, at best, the visa is authentic and belongs to the passport. How do I know the passport belongs to you and you're not just keep reusing it?
So from a compliance perspective, this is really the ultimate way to know, is the information you're using for compliance actually assigned to and belonging to the item you're looking at?
The same element that we talk about with blockchain. Blockchain did a phenomenal marketing job in convincing people blockchain protects your supply chain, and I'm sorry to say, that's BS. Blockchain is a phenomenal technology. We, so we sometimes use it ourselves with some of our customers that want it. It's a phenomenal technology for protecting digital data It does nothing for physical products. Why? Because as long as you rely on some sort of proxy to connect that physical product to your blockchain [00:45:00] data, you've done nothing. I could have a fake QR code or a co- a QR code that I moved from one item to another, connect to your blockchain, and now you're even overconfident. "Oh, look, my data's in the blockchain, then my product must be real." No, all it means is that your data hasn't been manipulated. It could still be a fake product or
a wrong product or a furbished product. So blockchain is great for data. Don't think that serves your compliance needs. You still need to make sure that physical item you're importing, exporting, showing compliance to, aviation regulatory authorities, healthcare regulatory authorities automotive health authorities, you still need to show that item is in fact that item. If I'm the FAA, I wanna know that the part you're installing on that aircraft, or if I'm the FDA, I wanna know that healthcare product you're selling is what you tell me it is, not because some piece of paper says so, but because the [00:46:00] product says so. That's a huge shift in the way you think about things, right?
Again, for people, governments have figured it out. For products, we're still relying on antiquated technology. So from a compliance perspective, it's the ultimate I'll even say the last technology you need because it's not a fancier QR code, a untamperable sticker a sticker that if you try to tear off will tear, a hologram that's four-dimensional with secret super dust and ink.
No, this is the actual product itself telling you what's going on. Stop wasting your money on all these proxy fancy additives. Just ask the product who it is
Matt: Great framing and something that comes up time and again, not just on our podcast, but in daily lives for manufacturers, the o- overall sen- sentiment of prove it, prove whatever it is that you're doing. And second, the physical layer, that I don't wanna say gets passed over in [00:47:00] all the discussion about artificial intelligence and things that we're doing.
A lot of that is data. A lot of that is, to your point, blockchain. It's, it's-- When you get to the physical level, there's not really that connection. And the point of the proxy, I think, is the heart of this, conversation, is that we do have a lot of proxies in life that don't help people prove it, and that's that's a big piece of the puzzle that I think's missing in the, especially the artificial intelligence discussion.
Roei: Not only do they not help you prove it, they create an overconfidence. There, there was a, a few weeks ago, maybe a month ago, I saw this post on LinkedIn that talked about a specific watch company, not our customer of course, a specific watch company that said, "We just came out with phenomenal protection for our customers. We have now this super-duper uncopyable hologram that's printed on our warranty card that comes with your watch." And that was like, and people were like, "Oh my God, that's amazing. Look at this holo..." And I thought, "Really? So [00:48:00] you have a proxy that's a hologram on another proxy that's a plastic card that goes in the same box as this product. Really? That's the protection? How do you know the card belongs to the product? Oh, look, th- there's a serial number etched behind it and printed on the card that tells you they belong to each other." Oh my God, that, that's like the... That's calling for any counterfeiter, "Hey, why don't you buy a set of those?
Now you have the real watch, and you have, a hologram that's on a card that goes with the watch." The, the it's proxy on proxy.
Matt:
Roei: you know that card is authentic. You don't know that the card belongs to the watch, except now I'm super confident about it, and I'm screwed. So the fallacy of these fancy things just doesn't work.
There was just this week in the news I don't re- it was somewhere here in the US, that the government or the, the state government was telling people, "Hey, watch your mailboxes, because there's been this spree of [00:49:00] mailbox theft where bad a- bad people are stealing mail with checks in them and then reusing those checks to deposit money." And how do they do that? They, they change the numbers and so on. The check is real, but they're, they keep changing it. And because the check itself is a proxy for money, right? Imagine if the actual piece of paper, that check, the piece of paper, was feature printed. Doesn't really matter what's written on it, but the piece of paper is feature printed. Now, let's say I'm the first actor, the, the legitimate actor, and before I even send you the check, I feature printed the check itself, and I say, in the original data, "This is a check made by," make up a bank, doesn't matter, "Roei Ganzarski Bank," and I'm now writing it out for $100. Great. Now you get the check. If let's say I'm the legitimate actor and I come to deposit it, the bank can take a picture of that check like they do today, but today, they only use the picture to read the print that's on the bottom, right? The bank account number, the routing number. [00:50:00] They don't actually read the rest of the check. But imagine when they were taking the picture, they were actually authenticating that check to say, "Yes, this is the check that Bank of Roei Ganzarski printed," and they added data to say it's $100 going to Matt. Imagine someone steals that check. Now take the bad actor. Someone steals the check or tries to reuse it. Now I'm the bank, I read the check, and I say, "Wait a minute. I don't know what's written on the check. I don't care what's written on the check. But this check, the actual piece of paper, is telling me I've already been used. I was issued by Roei Ganzarski Bank for $100 to Matt." The fact that it now says it's for $100,000 to Jim doesn't matter because the check itself, the paper, is telling me the truth about what it is. I don't care what's written on it. So you can protect everything and anything to a level that no proxy can do. So it really gives you that extra level of security on things that on day-to-day are so easy to manipulate. [00:51:00] Now that goes away
Matt: Pretty incredible to think about the level of proxy-to-proxy vulnerability that we have, overall. If, for our audience, if you're a manufacturer thinking about this, what's an action they could take over the next 30 days? And, if you were in their shoes, and as part of that, how can they find more about you?
Roei: Yeah. So even before that the way I first talk to p- to people about it is think about the consequence of getting your-- You're the manufacturer, the consequence of getting your product wrong, either unintentionally or intentionally. Unintentionally, I shipped the wrong item or I got the wrong item maybe there's a difference.
Or intentionally, someone's trying to dupe your customer. What's the consequence? If the consequence is, "Hey, I lost two bucks two dollars, two euros, and nobody really cares," then don't even contact us, right? Don't... We'll tell you, "Don't waste your money on us if there's no high consequence." Usually, we deal with only high-consequence items.
The consequence could be a lot of money, brand [00:52:00] trust
reputation. The consequence could be health, safety, national security. But there has to be a con-- If there's no consequence, I won't try to convince you to spend your money, and you shouldn't, right? At that point, the QR code, the barcode is enough if there's no consequence. So first of all, if there is a consequence, then think about what does that mean if you could solve that consequence, not by adding another machine on your production line that puts a sticker or another expensive laser etching machine that has to laser something on your product, or that you don't have to look at your product and say, "Where do I engineer some space to put now that data plate or that sticker or that QR code or that laser etching?"
Rather, I could make my product what it only needs to be because I could take a picture of it. If that's really something that's of interest, then reach out to us. You can just go to our website, alythion.com. You could write us an email, info@alythion.com and just, tell us that you wanna talk, and we'll gladly talk to you.
Usually we have a few steps that we do with customers or potential [00:53:00] customers. One, quick phone call. What's your use case? What are you thinking about? What are you doing now? Why are you calling us, et cetera. Then we'll say, "Hey, either ship us some of your products or we can come to you."
The-there's usually no cost to this because, it's really interesting. Companies really think that their product is unique, and many times it is. But from our perspective, a lot of times they're the same. So take, I have this watch, and this other company makes another watch. It's still, from the way we look at it, it's a piece of metal, maybe precious metal, maybe not.
Maybe it's a composite. Maybe it's, some sort of plastic and has a face, and it has a back. Both watches from our tech perspective are the same. But the watch company thinks it's unique, right? Or I make this part for the plane or this brake pad, and I make this brake pad. Oh, mine's different.
Yeah, of course it's different, but from our perspective, it's still a brake pad. We can take a picture of it. But customers wanna know, "Oh, great that it works with that. I wanna know that it works with mine." Fine, send me some of your products or we'll come to you. We'll test it. Assuming that it works, and usually it [00:54:00] does, then we'll go usually into some sort of, you know, trial effort.
Some wanna know, does it work in their production line? "Hey, I have a high-speed production line. I have a printer. I do things manually. Will it work in my production line with my products?" Great. Let's try it. We'll do some run, 30 days, whatever it is, s- number of products that you feel comfortable with.
And then once that's usually proven out, then we go into production with the customer. So very simple. The only thing you really add to your production line are cameras. It could be if you're only starting from the end of the production line. So let's say you don't care about in-line traceability.
That's fine. You want a finished product going out. You put in an industrial camera 'cause we're not-- we're assuming you're not gonna use a phone in your production line. You'll have an industrial camera. It works automatically through APIs and triggers. Product goes by, or maybe it's a manual process in QA, it's checked, put it under the camera, image is taken.
That's it. Immediately, it's ready for identification later on. Or if you wanna do in-line, put a camera where you think the critical parts are that you want to identify. As an example and this is [00:55:00] really important to note, we are not a QA system I can't tell you your product is out of spec or you have scratches on your product.
In fact, the question we asked before and answered is, what if there are scratches? Will it still identify? Yes. We are robust to a point we actually disregard the scratch and still tell you that's the product. So we're not a QA system, but how do some of the manufacturers use us in their QA kind of life s- cycle or program? Imagine you're doing in-plant traceability. You have multiple lines, multiple machines, and at the end you have a QA process, and you're noticing that some of the products, they're still within tolerance, but in some of the products they're starting to trend up in their tolerance. So it's supposed to be using that original example, plus or minus three hundredths of an inch, and you're noticing that some of these products are all at the two fifty, plus two fifty. Quality control, they're still within the band, but there's some tolerance. Imagine you could say every one of those parts, they [00:56:00] have one thing in common. They all came through machine number seven on line twelve. At least now you know that you don't have to try and figure out all of your lines. Go to that machine and try to recalibrate back to as close as nominal as possible so that you can fix that problem before it becomes a QA problem.
Or imagine you didn't do in line. Customer starts sending you products back and it's like, okay, what, what's going on with these products? Oh, these products, we can't tell you what the problem is, but we can tell you all these products you have all came through line number twelve, machine thirteen, second shift. Now you have some information you can go investigate and focus your investigation as opposed to what do I do? And think vice versa. Imagine you're feature printing every sheet metal that comes into your factory. Let's say you're using sheet metal. And you can register every one of those sheet metals, then identify them as they turn into various parts and get assigned to, let's say, a car. And now that sheet metal manufacturer calls you and says, "Hey, we had a bad batch. You really need to recall [00:57:00] those parts that came from that sheet metal." Now, a lot of times that sheet metal isn't serialized or isn't traceable. Imagine if you could know that the part that you part in that assembly came from that sheet metal that came from that manufacturer simply by taking its picture.
Now you know every single car assembly item to recall, as opposed to I'm recalling everything from the years of this and this or model that of that, Right. So even for recalls that no one likes, but sometimes you have to do, you could have them targeted because you traced in your factory every single part and what the end use was.
So that's how it's used for quality control, even though we are not a quality control technology. So that's kind of an element as well, and we can try that as we work with customers too.
Matt: Excellent. A lot of range there a lot of ability to prove it, and I think that's something that everybody is looking to find ways to offer that type of guarantee and that type of security overall. So Roei, thank you very much for joining us today. Great conversation, and I think it's one our [00:58:00] audience definitely needed to hear
Roei: Thank you. It was a pleasure, and thank you for shedding a light on this important topic
If you're on socials, give us a follow on LinkedIn, X, formerly Twitter, and Instagram. And if you're enjoying the podcast, please feel free to follow the show on Apple Podcasts, Spotify, or YouTube, and leave us a rating and don't forget to subscribe. Thank you again for joining us and learning more about how you can make your way.
Matt: Absolutely. It connects to something we come back to often on this show, manufacturing reshoring, and as supply chains get rebuilt domestically, the companies that win are the ones who can prove it, and that's exactly what Roei and Alitheon are building on the parts and materials side. To stay ahead of the curve and to help plan your strategy, please check out our website at www.veryableops.com and under the resources section titled Trump 2.0, where you can see the framework around upcoming policies from the administration and how they will impact you as a manufacturer.