Deploying an OEM AR Support Tool: The Value of Enabling End Customers

00:00 So I'm going to get started. Thank you for coming. Just in general, I'm going to keep this conversational and not as super formal. Right. I'm going to leave some time for questions, but I am Alexander Aulet. I’m the innovation manager for Harpak Ulma. Before going to Harpak Ulma, I was at PTC for about eight years. So what I'm talking to you today about is smart, connected machines and our digital transformation, and specifically about why it is valuable to put these tools in

00:35 the hands of your customers as an OEM. So if you're unfamiliar, right, OEMs manufacture everything from excavators like Caterpillar to Harpic. Alma does packaging machine but can be a spectrum. And generally margins are tight with OEMs. There's not a lot of fat built into the business. So that along with a very large customer base means efficiency is key. So really quickly, if a picture is worth a thousand words,

01:08 we have a bunch of videos here of the different solutions and types of things that Harpak does. So everything from flow wrapping, If you ever bought Twinkies, right, or Hostess cupcakes, it's an example of something flow wrapped tray ceiling, any kind of ground beef that you would buy. Thermoforming stretch wrapping, vertical bagging Cheerios. Right? So those are all food examples. A lot of medical projects in there over Covid , one of our largest projects.

01:40 And actually one of the first projects that we enabled with Vuforia was coronavirus vaccines. So working with a company that was contracted by the federal government, super cool project enabled by Vuforia, really , really cool. And that kind of gets at the heart of what I want to talk about today, which is we had real trouble getting texts into that facility to support them. So by giving them a tool like Vuforia or like augmented reality, they can support their equipment without us needing to go in. They still have the audit trail,

02:10 they still have all the benefits and it's cheaper and easier for everyone. So why digital transformation? So this is the digital twin of a piece of equipment that we have actually up in Canada. And it is obviously not sitting on our CEO's desk, but now it is so fundamentally why OEMs should adopt tools like Vuforia, like ThingWorx. The whole digital Transformation and Digital Thread suite is it allows us to do stuff better and faster, right?

02:42 And arguably a lot more scalably and a lot cheaper. If you figure if a customer has to go through a procedure like whatever, repairing a manifold on one of these machines prior to using tools like Vuforia that may have been we had to put a technician on a plane and fly them there. And there are economies of scale there where if 20 customers have that issue at the same time, either you're disappointing people or you're spending a lot more

03:13 on technicians than you need to. But if that can be an R procedure, everyone can do it at the same time with the same quality. The expert technicians can focus on other things that are maybe a little more ambiguous and ultimately it just delivers better outcomes Generally. I'll speak to this. This is an example of something that we do with our customers and kind of what this would look like. So imaginary customer Co is, for whatever reason, packaging steaks, syringes and salmon at the same time. But representative some of the different industry verticals,

03:45 we monitor the equipment with ThingWorx and there is a predictive monitoring aspect to it. So rather than saying even your machine is going to go down after 250,000 cycles and you're at 220,000, you should schedule maintenance. There's a condition based monitoring to this where effectively what it will do is look at the windows of what is your production schedule. So when is the next time you have scheduled downtime? What is the amount of time needed to do this procedure?

04:15 What are the parts that you need to do this procedure and when is someone available to support it, either internally or externally? So it can be an integration with your own ERP system or it can just be interfacing with a tool like ServiceMax that you can just schedule a technician from Harpercollins to come out, but effectively you get an alert, Hey, you're going to have this issue, you need to take care of it before this day in time. Pick a window when you would like to fix this. And ultimately, what they can do is go through,

04:45 check out the part or the assembly that they need, have it mailed to them, and it's all good to go. There may be some pre activities that they need to do that we can guide them through with Vuforia and then you know a much younger than me and devilishly handsome tech will show up and guide them through exactly what they need to do. You go through it, you verify it as well, using a tool like instruct in this case. Right. That's going to show that, hey,

05:15 not only were you guided through this step by step, but you did it correctly and the customer is back up and running. So in this particular video. There's a technician that's kind of doing it from Harpochloa, but this can just as easily be done with your own workforce where, Hey, we're just going to ship you the parts you need. Tell you the downtime window to execute it in and you're off and running. So the story here is powerful. We are definitely, I think, like every business, trying to dramatically grow our business and it's not just better for customers.

05:46 There is a fixed cost in terms of how many technicians you can hire to help you out. So if there's only so much technical labor that you can hire, you have to find other ways to scale your business to be more effective. What are the sort of things that we do in general? I'm going to run through them and then talk about a specific use case where we did this with a specific customer and kind of what the outcomes were. So, right. Why is R effective for customer self service? What kind of things do we do? So training is a big one, right?

06:18 Which we would consider anything that you're doing that is not on a production floor. So whether that be putting on a HoloLens and walking through how to do some procedure and a customer training center or in their own home, something like that, work instructions are real time walkthroughs. So right, you're guiding someone as they're actually doing the task in a production or whatever, setting inspection and checklists, obviously. Or maybe if you're unaware of the industry, very popular among our medical customers,

06:49 but also really important for some other use cases. So generally I would say you want to use R for things that are complicated, but sometimes R is really effective in doing relatively straightforward things that are extremely important to always be done correctly, right? So an example of this is our customer, Maple Leaf Foods uses these inspection and compliance checklists

07:19 for sanitization of their equipment. Right? No one wants a food recall. Certainly no one wants to accidentally kill someone at work, which is a risk in CPG. So what they do is they use the expert capture platform because it has insights, because you can require photos, because you have an audit trail to basically say, was this equipment sanitized? You know, did they do a 30 minute procedure in 30s then it probably wasn't done correctly, stuff like that, to make sure that they are providing safe product to

07:52 their customers. Remote collaboration, very simple, right? A tool like chalk where rather than having someone hop on a plane to debug something and come out and solve it, you can use just video chat with R for lack of a better term. We went through a journey with this in terms of a self-service tool for our customers, where initially there was kind of a conversation about like, well, if they have my cell phone number, why don't they just call me? And then we had one specific customer support case right at the beginning of Covid,

08:26 where we spent, I think 5.5 hours on the phone with a customer, ran through everything. And step number two was, is this connector plugged in all the way? And they said yes. And the minute they got on shock, it was not plugged in all the way. So really simple problem. But if you're just, you know, kind of playing the trust without the verify game, it can lead to problems. And then the last piece is performance monitoring and analytics, which is powered by ThingWorx. What does this all look like? Again, this is an example of what something might look like training wise,

08:57 where you're overlaying R on top of an actual machine guiding them through how to take a tool out. We do this across kind of all our product families, so everything from really, really simple machines that would like seal an individual sandwich pack for lunch up to automation projects that are the size of several football fields in complexity. So a lot of this can be simple, it can be complex. A lot of the value in this is that people, especially a lot of frontline workers,

09:31 just by their own preference and learning, are experiential. So rather than having them watch a video, if before they start their first day or on their first day without, you know, shutting down production and taking them out to the floor to see the machine, if they can put on a HoloLens, they can take out a tablet, walk around the machine, go through the procedure, already have experienced it before. They truly experience it can be really powerful and you can do it at any time, anywhere without paying for trainers necessarily. Next piece, work instructions, right?

10:03 I think we have been the most happy and our customers have been the most happy with this sort of solution. So that's our individual, you know, little sandwich machine, for lack of a better term. It's used for a lot of other things, but that's kind of what you're seeing, a relatively small package there in an. Expensive machine, but you still need to train people on how to use them, Right? And at that cost point, you know, if you're spending 100, 150 grand on a machine, do you want to spend 20 grand on tech visit? Probably not.

10:36 So what this work instruction is doing is an expert capture, showing them how to set up the machine for the first time. Right. So we can also localize this to any language. We can share it with our partners. Really easy stuff. Not a high barrier to entry as an OEM, right? Any sort of technical training organization can author this stuff. And ultimately what it means is for them spending a week, two weeks at most on building out this training content. Now, we never need to do in-person training for this machine family again

11:09 or just minor updates. So really, really valuable type stuff. This is actually a demo that if you go down on the floor, you can actually see and do this yourself. So this is a manifold repair and replacement. Another example of work instructions or inspection and compliance checklists. Big piece on this is that step check the AI functionality is enabled and this has been extremely effective with our customers.

11:39 So you can see, right, you don't necessarily even need to get their feedback about is this good, is this bad? Right. It automatically is detecting is this in the right state? Is this in the wrong state? So one extremely effective at training people and guiding them to the correct outcome. But two, it is an added layer of trust and safety, I think built into the process. Right. We do this with all sorts of things like lockout tagout for our customers and for our machines so that not that you want this to be the final stop,

12:12 but it's a layers of security and layers of redundancy built on top of each other. And as a benefit, you get a really nice audit trail of how your equipment is performing, what maintenance needs to be done on it, how all those sorts of things. So I would encourage you to go down to the floor and check out this demo over by our machine if you have the chance, because it's fun. Next piece, Remote Collaboration Assistance. Like I said, we use chalk. Really effective enables customers basically if they want a tech there yesterday,

12:46 they can get the next best thing. So that's all I would summarize there. Asset monitoring and analytics. The big piece that I would say about this pertinent to this conversation is that effectively ThingWorx and similar solutions are similar Iot capabilities can actually be really, really effective in how you target your AR. So you'll see at the bottom there I have a top ten fault list.

13:16 So fault number one and two are basically the machine is running so fast they can't feed it product by hand, so they just wave their hand in front of a photo to slow it down. But the rest of those fault reasons, basically how we prioritized what AR content we're going to develop was, okay, for instance, you're spending 20 minutes loading film and that should take five minutes. You can use the Iot to tell you that you need to develop training content or you need to better train people for those procedures.

13:47 So we use it very much not just as a continuous improvement tool, but almost as a researcher or a data gathering tool for where opportunities for AR and where opportunities to better educate our customers are right. If we get a lot of vacuum system failures in our equipment because they're not doing the maintenance that the prescribed interval, we can update the interval, we can give them better training, we can figure out why they're not doing that. It can really be a good data gathering tool for us and our customers.

14:18 So I'm going to move on to a specific example in a second. We're really quickly, you know, as an OEM, what are the challenges that we kind of face rolling out these programs with our customers? A big one is bureaucracy. So particularly in a production setting, we run into a lot of it and OT hating each other. So one owns all the general IT Wi-Fi, what have you, systems. The other one owns everything on the actual operations floor.

14:48 And if the operations floor wants to connect a new piece of hardware to a network or do something like that, it can basically devolve into tribal warfare. So not a great situation. General cybersecurity concerns right beyond the normal. You're putting devices in our plant. The other piece is you're putting devices with lots of cameras and video recording in our plant. So not necessarily the funnest conversations to have all the time. Skills gap is a big one, right?

15:18 So if you have an organization that in some cases are billion dollar companies that have an IT group of two people, not setting yourself up for success, and then kind of to go back to the tribal thinking it's not always necessarily fighting between each other. Sometimes there is just a general cultural resistance to change. So an example of this that I never thought

15:48 I would have to deal with was at one customer. The workers that were wearing hollow lenses were getting bullied by their coworkers because they thought they looked goofy and had to have a conversation with HR about you cannot bully people for wearing hollow lenses, which it's great. So at least the job is always interesting. But examples of some of the stuff that we've run across in terms of how to solve those, right? If you have executives that know where they want to go or just leaders in general at any level that know where they need to go

16:19 and can articulate where the value is in this, it's a big thing. It goes back to engagement. It goes to having a culture of improvement. I think anyone that wants to go through knowing that something is 100% going to work off the bat is kind of setting themselves up for failure. You're always going to make mistakes. You're going to run into challenges. That's just part of adapting and improving. I think the biggest thing here is showing shared value, right? So if you show that, hey, you can have your maintenance people deal with your backlog

16:53 because they don't have to work on X, Y, Z, because you're equipping other people with R if they don't view it as a threat to their job or if you can get them to see that it will ultimately make them look better or help their company make money. We deal with some customers that have a profit share, right? And 1% of OE on one of our lines for this particular customer is roughly $1.8 million. And if they hit their targets, everyone gets a bonus, a pretty great one.

17:25 And it's obviously tranched. But every. Everyone on that plant floor has a financial stake in minimizing downtime. So if you can say, Hey, I can take a procedure that takes you a week and get it down to a day, they're going to get their bonus at the end of the year. There's a financial incentive for them to do it or from a manager perspective, you look better to your boss. You get praise and accolades and get promoted and then have more money to buy our equipment . So it works out for everyone. So really quickly. And then I'm going to move

17:56 on to my specific example of one thing that we did and how it was effective. So why OEMs are really in the best position to help out our customers with this stuff? A lot of it has to do with proprietary data and just knowledge of the equipment. We know how to service it best. We know how to build it. We have all the 3D CAD data. We have all sorts of very niche and specific knowledge to our equipment that customers are not going to do or as much. I love our systems integration partners.

18:26 The process for them is going to be to ask us for the data and we're going to say no. So OEMs are uniquely positioned to be able to help their customers and I think they have a responsibility to do so. We have customers that now specifically spell out as of I think actually three years ago, that they want work instructions provided in Vuforia or they will not buy the equipment. So the customers are getting smarter and smarter and driving in that direction

18:57 and OEMs will either adapt or die. It's as simple as that. But we do have the most knowledge on our equipment, so we are the best equipped to develop content for it. So talking about a specific example, right, of ah, self service and what kind of the results were and kind of how we move through it. So I'm going to talk about the tool rebuild for our mondini line of products. So basically what that machine is, I'm going to skip forward a slide really quickly.

19:28 Typically very large, relatively expensive and complex equipment. So it's a tray sealer. Effectively trays with product go in, tool comes down, seals a film over the top, items come out the other side. You can think of examples of products that you would see go through. This Lunchables are a really good example, right? Plastic tray sealed film on top. Actually a surprising amount of medical content as well. Ground turkey, ground beef, all sorts of stuff.

19:59 But generally these machines are valuable because they run very, very fast. So you can effectively on a single line if it's running up to speed, run millions and millions of products through in a year, that particular tool in the middle needs to be taken apart and serviced somewhere between every 500,000 and a million cycles, depending on how you're running the machine. It is very, very complex. There are over 20 different individual procedures that need to be done with

20:32 a total of around 17 to 1800 unique steps. So doing that from memory, pretty challenging on the very best of days. So if you combine that with the issues that our customers face that are, you know, all our customers are dealing with kind of three common issues, so they can't hire enough people, period. The people that they can hire are not necessarily as technical as they would like.

21:03 And the third is that they have very high amounts of turnover. So so they're increasing their tech debt. They can't hire people, period. And the people they can hire leave really quickly. And a lot of that is just that. Working at our customer sites are not necessarily fun places to work. They're cold, they're wet, they don't pay super well. It's very understandable that people don't necessarily want to work there, but there's still challenges for their businesses. So ultimately, we have a quote here from one of our service managers that basically

21:36 we view this as an existential problem for our business as well, and that there's only so much you can automate and there's only so many situations and that you can automate. A lot of times companies are now automating their processes not because they want to eliminate jobs, but they can't hire for the jobs that they do have. We have some customers that have a quarterly turnover rate of 40% , so they have to hire 120% of their entire workforce a year just to break even.

22:08 It's extremely challenging. So where does Vuforia come in? This is an example of the actual tool that we're using. So what we did is we went through the tool rebuild and built out the entire tool rebuild in augmented reality. So using Vuforia, they're guided through every individual step, every individual process that they need to go through to do this particular thing. And as you can see, step check verifies every single thing along the way. So you know why we chose this particular thing to focus on is one,

22:44 it is the single largest pain point for our customers. It has a huge amount of value. Their average savings in terms of implementing this is they save all in about a quarter of $1 million per line and save downtime in saved labor training, cost lost production. So very, very high value. It's obviously extremely complex, but the other benefit is it's repeatable. All they have to do this anywhere from 4 to 8 times a year and pretty much our entire customer install base

23:17 for this line of machines has to do it. So a lot of value, it makes sense to do with AR and it's repeatable. So taking that into account, kind of what that ends up looking like is we have these procedures when they go to buy the machine, we sell it with them and then they are basically able to either before the machine is ever installed, have their maintenance staff come out to our facility and run through a tool rebuild using this tool.

23:48 So they're familiar with Vuforia, they're familiar with the headsets. You know, they're able to do it and they have that familiarity. And then they can be the super users that help other people. But it also just helps us find any problems and gets them, you know, abuzz about it. So moving on. That's just a customer quote. Basically talking about that again is Maple Leaf Foods. It's from their maintenance manager. They're basically talking about how with augmented reality,

24:19 despite their labor shortages, it's cheap, it's faster, and it's cost effective to train people with. AR. So how long did this take us? Who developed it? Basically, we worked with a partner from the outset and identified with several partners what their major pain points were. That's how we arrived at this. Overall, it was an iterative process. I'd say we probably went through 5 to 10 different iterations of the tool rebuild, both updating as Vuforia became more complex and more powerful,

24:51 but also as their requirements changed. Where, you know, originally, for instance, like if there's three o rings on the bottom, there might have been a step that said Replace all three O rings. And then we found that people still were skipping some. They weren't doing them right. So understanding that we really needed to be more granular and add more verification controls on it. Overall, the total project in terms of initial R&D was probably about a year, but now we can develop, you know, for every tool we sell,

25:23 I would say it takes us about a week to customize and deploy it. And so ultimately this is another quote from Kraft Heinz. Basically, you know, they are one of our customers that for the facility where you piloted this at, they have 12 lines that run 20 hours a day, 364 days a year. So they rebuild a tool about once every two weeks across all those different lines.

25:54 So the end result of that for them and this is kind of reflected across all our customers is we were able to take something that takes a week for them to do and get it down to where they can do it in a single eight hour shift. So to date, and I would knock on wood if I had any here, they have not had an error in terms of going through it, and that has been with no prior training on this task. So additionally, we can deploy it to anyone anywhere if they buy a similar machine

26:28 and stick it in a whole different geography, we can get them just as effective pretty much immediately by putting this tool in their hands. And for us, you know, this has been a huge competitive advantage. So we have 100% sold machines that we would not otherwise have sold because we offer these solutions. Because again, across our industry, everyone is dealing with labor shortages, Everyone is dealing with skill gaps, everyone is dealing with turnover and taking all those things,

26:59 not necessarily in a vacuum, but if you understand those, it is a way to make basically your new hires as effective as your most expert workforce in the procedures that you put it in place. It's not foolproof. But an example of this is we took people who not necessarily even were in maintenance or operations jobs, but totally non-technical people. We took our invoicing manager at halfback and had her go through this procedure.

27:30 I don't think she had necessarily used an impact driver before and she was able to rebuild the tool and then put it in the machine and have it run correctly for a customer demo. So that was kind of when we realized, Oh, like this might be a really, really effective tool. But ultimately, I mean, the end result is it's just a better way to support our customers, right? They don't have to deal with aside from like the fact they can do it faster and without errors, which is again worth potentially millions of dollars to them.

28:00 It actually is easier for us because rather than running instructor led trainings, every time we sell a piece of equipment, we can put Vuforia in their hands and just make them again, as effective as our most senior technicians. Well, thank you, everyone. I appreciate it.