Inside the Digital Shift at Ireco: A Candid Client Testimonial

[00:00:00] Filip Popov: Hello and welcome to another episode of Espresso 4.0. Today we're joined by a very special guest, Xavier Simon. Xavier, welcome.

[00:00:10] Xavier Simon: Thank you.

[00:00:13] Filip Popov: Xavier, why don't you start us off with the basics? Why don't you tell us a little bit about who you are what's your role at Ireco Trading and Production?

[00:00:27] Xavier Simon: Thank you, Philip. So, my name is Xavier and, uh, I've been working for Ireco for four years as a plant manager. Um, prior to this, I made my study in engineering and in finance. I worked long years for Caterpillar, uh, mainly in Europe, but also in Japan. Then I moved back to Luxembourg and I worked for Goodyear, then Cebi, and then in 2021, I joined the Ireco team, uh, as the responsibility for all production in Luxembourg and Italy.

[00:01:08] Filip Popov: Excellent. I mean, a colorful background, both in terms of, uh, companies and geography. You probably learned a lot from those experiences and had brushed up with, uh, different digital transformation, you know, efforts and projects. So, having said that, uh, digital transformation is often associated with like big flashy technologies, but the real impact happens in day to day operations.

Um, what's one aspect of Industry 4.0 that you think is underrated, but has made a real difference in your experience?

[00:01:44] Xavier Simon: For me, the first is the people that are installing or using the tools, their knowledge. That's the first, uh, because you can buy a 4.0 on-hand system, uh, but it doesn't mean that it will work. So, the knowledge of the people that are creating the tool and using the tool is very, very important. The second is, uh, you can have two situations.

You use a tool that you will customize to 20 percent and use 80 percent that is like everybody else. Or you can do the opposite. Um, I've done both and, uh, they have both pros and cons. Uh, unfortunately, I would say for our parts now, our system is quite very, uh, unflexible and so the, uh, restriction that we have with food, uh, does not allow us to a lot of discrepancies and the request of our customer makes us very special on these needs.

So, that's why we needed a kind of tailor-made solutions, uh, to improve our machines. Because I am not a large industry. So, we purchase different types of tools, different types of machines. It would have been easier if all machines that I have are from the same maker and producer. And then I would connect using this.

Unfortunately, it's not the case. So, using one will not fit all because every machine has a different connections and, uh, it's part of the great job that was done by the team, I would say, connecting to different machines, the data, analyzing the data and see what is relevant for us or not.

[00:03:39] Filip Popov: Gotcha. So, I suppose if I can distill what you're saying in terms of what the underrated things about Industry 4.0. One is people, right? They need to take ownership of any type of technology that comes on board and they need to receive upskilling and training in that technology. And secondly, the way we go about technology and Industry 4.0, whether making a very important choice of whether make it in house or get it off the shelf.

[00:04:06] Xavier Simon: Yes.

[00:04:06] Filip Popov: Each of those coming, coming with a set of pros and cons. Okay, yeah, absolutely. Fair enough. Those are, those are very important. And additionally, you said one of the considerations that one must take when deciding on which approach to go with. Apart from obviously knowing what their internal resources are to develop technology is the fact who their supplier of their equipments are right and not being vendor locked into one supplier solution as opposed to having a solution that can fit and integrate, um, other machines and data regardless of who produced it and who you bought it from.

[00:04:47] Xavier Simon: Yes.

[00:04:47] Filip Popov: Those are very fair considerations. Indeed. You've mentioned food industry, obviously, because that's what Ireco's main, uh, main, uh, that's what Ireco's industry is. You guys are a dry foods manufacturer and particular pistachio, which is if you noticed why I've decided to come themed in a green sweater.

Having said that, the food industry has to balance efficiency, quality and sustainability all at once, right? So, do you see digitalization as a competitive advantage or is it becoming a necessity for survival?

[00:05:22] Xavier Simon: I would say, for the years, the previous year, I would say there was an advantage that we saw, but now, um, it's, um, it's a necessity. Uh, let me explain. Soon, uh, the CSRD will be, uh, something that it's needed, required for major companies in all over Europe. And so, sustainability will not be something that you can do nicely and prompt as nice. It's something that is embedded in the laws. So, we need to improve. For us, of course, we have like every large company, uh, many teams that we can improve. But, uh, as we roast, uh, nuts, we use quite a certain amount of gas and, uh, using technology to improve that consumptions makes us more sustainable.

So, that's what one of the team that we purchased or pursue last year. And that's, uh, where, uh, I would say your team help us a lot. That's why we are focusing, still focusing on because the other parameters are more, uh, out of our heads. That means where we need to purchase. Well, it depends on regulations.

Uh, the type of, uh, containment or packaging or anything is defined by our, mainly defined by our customer. We can provide solutions, but the end decision is always the end customer. So, depending on where they are in the sustainability evolution, then you can see a difference. On our side is reducing consumption of water, electricity, gas. And gas is, I would say, the one who's the most important in terms of CO2 production.

[00:07:17] Filip Popov: Yeah, I tend to agree with you. I don't think it's only in a competitive advantage anymore. I think it is in fact, a life, a lifeline and a way to survive within the, within the current, uh, landscape, industrial landscape. To kind of go back a little bit more to the point and hammer in on the point that you've mentioned regarding people and the importance of it, right? Um, when talking about technology, technology is really only half the battle. Adoption is the real challenge. So, what's the hardest part about getting teams to embrace digital tools and how do you overcome it?

[00:07:56] Xavier Simon: Well, you know, on our side is, uh, we have a lot of, uh, work that is done on knowledge and, uh, 53 years of knowledge in the company. And then when we have a technology solution that gives us a different answer, it's really difficult to explain to people that it might be the right one, even if we have done always the opposites.

You explain, you can explain it, use, I would say Kodak solution saying, yeah, you know, Kodak used to make films for cameras. Well, they don't exist anymore. So, it's, uh, people understand that. They listen to this, but, they always say, well, it's not applying to us. In fact, it is. And so, uh, using technology to improve even by the slightest numbers is really important.

And then the way we have done it, uh, it's of course by discussion, training, but also by proving. So, um, it's always good when a person will challenge the system and say, okay, I'll make it better than the system. And then when the system gives you an answer, you play with it, get the best results. And then the person that has tried,

is failing the system. They say, oh yeah, my product is okay, but all the parameters are not okay because humanly it's very difficult. You can optimize one, two setups, but you cannot optimize 20. It's not possible. It's how our brain is like this. So, that's what a machine or an AI can do much better than us and provides, I would say different setups that we can use and using then our knowledge based on those setups to find the best solution for us. And this gave us the best results that we have. So, but it's more about challenging people and then, uh, giving them the right information.

[00:09:56] Filip Popov: Just a little bit off of that topic. Um, you've worked in different organizations at different levels. Right now, at Ireco, are the COO and head of production, so you can tell people to do whatever you want. Right? But you've also been in positions where you had people that are above you naturally, right? Where do you think, do you think that these types of changes and adoptions of new technologies need to come from top to bottom, or is there a space from bottom up or mid up?

[00:10:32] Xavier Simon: It will, well, it depends. I would say for one solution, if you have one problem, uh, then it can come from bottom up. Because it's working in Japan. That's what they call Kaizen. Okay, that's improving step by step and you can do it with systems. Uh, of course, uh, if you want to make it at scale or company scales, people from the top must be convinced.

If they say, okay, you apply it, but I'm not convinced, I'm not using it. It will not work. I can give you an example. For us, uh, for long years, uh, IT was not something that we lacked. It's something that we used or we needed to have. And, um, then when you see the hierarchy not using the tools that you have to use, the people will tell you, well, why should I use this?

And you're not using it. So, you're not, you have to be the first, I would say, advocate of the tools. And then of course you have to try and so on, but you have to be the first advocate. So, I would say the main solutions will always, from a company point of view, come from top to bottom. Of course, improvements using the tools can be up and up from bottom to up.

[00:11:54] Filip Popov: Yep.

[00:11:54] Xavier Simon: And then, when you use the tool, it's like everything, you can improve it daily. So yes, bottom up will work, but when you need to start, it's purely a management decision.

[00:12:05] Filip Popov: Gotcha. Okay. So, looking five years ahead, do you think AI and automation will fundamentally change the way food production operates or will human expertise always be at the, at the core?

[00:12:18] Xavier Simon: That's a big question. Um, honestly speaking, I would say I'm expecting that in five to ten years, not only production, but a lot of stuff will change due to AI. The way, uh, let me explain our system. The learning curve of our system on our roaster, for example, is quite huge. So, I'm expecting that the roaster itself could in the next few months be fully automatic.

So, the AI will drive 99 percent of the roster. We might request a human intervention in certain cases, but we will define those cases. So, it means the intervention of a human will be lower, but more important. So, like a lot of people will tell you yeah, it will kill jobs. I think it will change the jobs. Some jobs, of course in production will disappear, but some jobs will be created by this because you need more analytical solutions.

You need more people with practical, uh, expertise, but you will still need experts. So, somebody that knows the products and that person needs to grow in terms of system. I guess you, he will work less, with, uh, the tooling and, uh, so on that today, but, uh, he will use more his brain to take rather higher decisions.

So, yes, i'm expecting that AI will change a lot in production in the next five to ten years.

[00:13:51] Filip Popov: Fair enough. I think where you add is that it will change a lot. Humans will still play an important role, but AI is likely to take over a lot of the repetitive jobs that require a lot of different variables to be taken into consideration at the same time. Right?

[00:14:12] Xavier Simon: Yes.

[00:14:12] Filip Popov: To make better decisions. Okay.

[00:14:14] Xavier Simon: And a lot of administrative, administration jobs can be exchanged because for example, for traceability in food, for the moment we have, we have to scan, we have to do this. An AI can do that from the first step to the end step at once. So, uh, it's, uh, those scans, it's not, it will not change everything.

The job will stay the same. And I'm hoping that humans will have still places in this, but it's, uh, frankly speaking, I think using AI to help us making better decisions, be more efficient in using as a great tool in the future. I would say for me, AI is, uh, kind of industry 5.0 because it will, uh, use the automations, the knowledge, but will provide solutions or optimizations and provide you with choices that you can make.

[00:15:14] Filip Popov: Gotcha, okay. Now on that topic, if we could stick around a little bit. Naturally, you are one of Wizata's clients, one of our favorite and long lasting clients and you are also here in Luxembourg, just located as we are, um, having said that before you have decided to actually go with Wizata, what was the biggest obstacle you were facing before starting the search for an AI solution, AI technology?

[00:15:46] Xavier Simon: Well, it's, we are in Luxembourg, but we are, we consider ourselves as a small company because, uh, well, we are, the number of people here is quite low, even if we produce a lot. And, uh, in terms of money wise, it depends on, uh, the product that we, uh, we do. We're not a large company. So, the first, so, uh, and then the second is we don't have a large knowledge about IT And especially AI. So, the first point was to find a company that can support us with that kind of knowledge and be adaptable to us.

It means understanding our own processes before saying okay, we'll sell you this solution. It will fit.

So, what we, uh, the company that we had, or we discussed with, and we thought that is what was on one of the, them on the, uh, the start was the way the person of the contact entered and said, okay, but before discussing any solutions that we would fit before discussing anything, we looked at the interest of the company to us.

That was the first point. The second was, uh, of course, uh, it's a huge amount of money. Uh, it's important, uh, to make and make that choice was very, uh, very disruptive because we didn't have any experience. So, uh, having the Lux innovations behind, because we're all based in Luxembourg, so we all met by Lux innovations and, uh, that support was quite great because they provide us with kind of an amount of companies that they had experience with.

And knows that it would work. So, the second point that we took in, I would say in conditions that saying that even if Lux innovation is giving us a stamp saying, okay, this, those companies are, we have experienced and he works with, and that creates a lot. And, um, then the last point was the discussion in the bus.

So, yes, still under Lux innovations, a part, but, uh, being able to discuss our problem and what kind of type solution I would say you have worked. Then give us a, a kind of, okay, create a bound. That's the first thing.

The second was also, okay, we had mutual interest for this to work. That's the first point.

So, it's not a case show, you take care of what we wanted and, uh, what we needed for the future. And then the last of course was the, uh, the price offer. At the end, we were, three of, only three companies were matching what we wanted. And then, of course, your proposal was, uh, honestly, one of the best, well it was the best that we received, so that's how we went to it.

[00:18:47] Filip Popov: Yes, thank you. Well, that's a little tap on the shoulder for me there, but, um, but aside from, from that, can you, can you expand a little bit about what the problem was, the business problem for Ireco that you wanted, that you sought out artificial intelligence to actually solve. What was the opportunity in the business that might not have been a problem. It might have been an opportunity that you saw, you know, like, yeah, I think we need an AI solution that can help us solve this.

[00:19:13] Xavier Simon: So, uh, we had two factors. The first is currently our machines are fully, fully managed manually. And the second is, uh, we react to the proportion. So, uh, as we are in the nuts industry, we, it's an organic, nuts. And then, uh, parameters are changing as it is natural. The way we react is, uh, I would say we change parameters every two hours or every hour depending on the productions.

But this matching the fact that we purchased a new roaster that was able to produce much more with much less while the reaction time needed to be faster. And then humanly we could not. So, that's the first point. The second is, um, well, uh, the crisis in Ukraine, uh, put a huge pressure on gas costs.

So, for us, it tripled our perspective on gas consumptions costs. So, we needed to find a way to reduce this. Last point was also where the CSRD is coming. So we, we need to be, uh, to reduce our CO2, uh, footprint. As I said earlier, it's, gas is one of the only way I have on my hands to manage. So, it's the only way I can, uh, reduce my CO2.

Of course I can reduce, uh, asking the boats, uh, to transport, uh, the containers to be electric and so on, but it will not, it's not my decision. Okay? And then it will take a huge amount of time, but this I can. Using the gas consumption, something that we could work on. And so, um, I would say the challenge of the machines, because it was new machines with much more parameters to manage.

The second was the timeframe, so the time to produce, to react to the machines. And the last one was really to be, to optimize, uh, gas consumptions. Uh, that's the three facts that pushed us to move forward with AI.

[00:21:23] Filip Popov: Gotcha. Okay. Great. And how has Wizata helped you make a real impact, uh, solving these problems?

[00:21:32] Xavier Simon: Um, the first, let me say, let me, let me explain. Uh, so I will give names, it's Jeremy and Albert. And, uh, the team ask us what was important for us. Then, uh, the whole team of Wizata was available and came to us, to visit the plant to understand what we have done. And then they looked clearly at where to connect and what it didn't mean as a data.

So, the first step was, okay, we take care of the connections. Of course, I had to work. But basically, they helped me to understand, okay, I need to connect from this to this. And this is how we can do it. They kindly helped me. Okay. I've done some work, but the direction was quite large enough to, for me to fulfill what they needed.

The second was the, uh, we went to the data analysis and said, okay, we have those parameters. What are the most important? What are the range? What are the reason? So, it's making the data compared to the reality. And then learning. So, it means, okay, we have this data, we make those certain points and then learning and then having a weekly feedback.

So, that was for us very important because maybe the meeting was taking 10 minutes in a week, but sometimes it took an hour because we found something else. So, uh, and then proposing of solutions. So, uh, our knowledge and our working on the machine, we know the products and we know our machines. How to connect, how to improve and then all the calculations, I would say Jeremy and Albert's team, uh, made a great job on this. So, uh, currently, uh, we are still using what Wizata provide, we thought or say, okay, can, instead of using one roaster or two roasters or three roasters, how can I, by ourselves, how can we, uh, I would say manage differently, not using all the roasters, but cutting roasters.

That was our first options, but we didn't think about this before. Okay. You know, the discussions then, uh, but that was out of your work, but your work helped us to think differently. And then the second was, okay, gas was different, uh, pressure and work with the CPM. So the producer, the producer of the roaster, uh, based on the data that we found, we had a feedback, uh, from, uh, CPM's team and Ian's team for seeing what we can do, changing some parameters into the machines.

So, that was the second part. And the last was, okay, uh, knowing and learning from our setup points. The tool provided by Wizata gave us the parameters that we play with, give us a setup of trial, and then we make trials, and then we see that we could reduce. Using those four, uh, sections in gas for the same volume, we know that in 2024, we reduced by 16.7% the gas consumption for one ton. So, it's not only the Wizata tool, it's all the steps that we have made that can reduce this. And Wizata, not only by using the tool, but by sometimes asking the questions out of the box, made us think differently.

[00:25:06] Filip Popov: Yeah. Yeah. I think, and you can correct me if I'm wrong, the summary of that would be first and foremost, the approach of our team in customer success. In first, in, firstly in, kind of collaborative workshopping, the idea of what the use case is. Um, then working with you together collaboratively to look at the data, combining your guys' expertise in producing nuts and understanding your machines and our expertise in knowing and interpreting data to find what are the most relevant points.

And in that process of discovery, you also figure out other opportunities for business and improvements, both also not only in the data itself, but also in terms of how you manage your operations. Uh, example of using one roaster as opposed to the other all three and last, but not least using Wizata as a tool, basically to house the results of the models and calculations that we've made from the data, that you use today as recommendations, to manage the productions.

[00:26:15] Xavier Simon: Yeah, I would just add one point, is availability. So, I would say anytime I need something or I have a question either by email or either by phone, I always have an answer during the day. So, that was, that was very important for us. Because we produce quite a lot of tons and, uh, I would say the availability of the team was quite great and still today, so...

[00:26:46] Filip Popov: And you can count on it going forward. So, don't worry. Um, okay. A little bit about the tool. Not all digital tools are created equal. So, what's the one feature of Wizata that stands out the most for you, for your needs?

[00:27:01] Xavier Simon: I would, I would say two points. The first is, it's really easy, easy, easy to use. It's user friendly. Uh... Jeremy gave me, gave me an hour, an hour and a half of training, but not to use the tool, but to explain how it works. And then based on this, uh, I was able to create my own digital twins, my old connections.

I was able to use, to do my own charters. That's the first thing. And then the second is, uh, if I use it on one, it's exactly the same to use on all the, I don't have to adapt on all the other tools. So, if I know that in the future, if I use it on a packing machines, then I have the same scheme, I have the same setups, I have the same things that is, uh, for the roaster.

Of course, the pinpoints will be different, but as long as I understand the, uh, the structure, I can create my own graphs, my own, uh, reports. And it's easy to measure. The second is, uh, everything is paramet, all the parameters I can choose. I don't need to send you an email saying, uh, sorry, speed, uh, I need, uh, to change from one minute to every ten minutes, uh, because the graph is, uh, something.

I need to see something. It's, uh, it's my own choice. So, it's my basically, yes, the frame is yours and, uh, but the rest is everything that is in, in this is mine. So, I know that in my company for the moment we have, uh, I'm using it, but it's, I would have no problem that, um, somebody from my team uses it on the screen and he's created his own reports and his own visualizations quite rapidly.

So, that's the two things I would about the tool.

Okay. Perfect. Thank you. So, success is measured in results. You've mentioned some of them. Maybe you can mention them again or if you have any other ones. Have you seen any tangible ROI since implementing Wizata?

Uh, so for the moment, we honestly, as I already told, we didn't make the specific on the Wizata points, but we have seen that, uh, our results using all the aspects that we have. We know that we have reduced the consumption of gas by 16.7%. Our goal or aim for 2024 was a bold goal of 15%. And we said, yeah, we might achieve it.

We didn't expect to achieve it at the first year. This is the first, it's not only the two, as I said, it's not only the result of the tool itself, but it's the way we started to ask the questions and then, using the tool to make us better. So yes, we have made improvements. The second, um, point I would take for my company is that, uh, we moved from, I would say 20th century to 21st century because we moved from normal IT usage to AI usage.

Uh, and it's a small amount. But this help also, uh, some people to use different tools in different areas that we would never afford. So, it changed a little bit of the aspect of the IT vision of the company. Um, and then that's something that we want to develop.

[00:30:35] Filip Popov: Uh, if I'm not mistaken, there was also some results that you, um, had in terms of increasing yield of the production from the roaster. Can you share those success as well?

[00:30:45] Xavier Simon: Yes. Yeah, so the, we, uh, increased basically all the, let me, just to avoid numbers, all the aspects that we wanted to achieve, we achieved it. We were expecting to do it in two years. We have done it in one year. So, uh, by using, and I don't know if Wizata is giving the same, uh, support as everybody, but on this, for us, that was, and it is still marvelous because it helped us quite a lot to understand a lot of things.

Uh, yes. Uh, using the tool is something that gave us a different strategy on what we can expect and, uh, but the team was much better. I would say if you take the number of hours of discussions, uh, of the whole team, I would say from you to your own CEO, to your COO to your data scientist, to all, they all gave a pass to better integrations and results in terms of ROI on what we have done.

[00:31:59] Filip Popov: It takes a village to raise a child, right? And, uh, that's how we approach also our projects. Every project is a little child. So, um, choosing the right technology is a big decision, obviously. And you took also quite a cautious approach as well, as you've mentioned earlier. What advice would you give to someone considering Wizata?

[00:32:24] Xavier Simon: Give them a call. So no, it's, I'll be honest, uh, of course, working with Wizata was quite easy for us. And, uh, as now we know the people of the Wizata team, I think it would be the same for everybody. Uh, the way you approach a project, uh, makes it easy. Uh, and I don't think it would be different if we produce nuts or if we produce tires or produce, uh, concrete or whatever in terms of the industry. Um, I wouldn't even expect a difference.

[00:33:03] Filip Popov: Although it does help that you offer us nuts every once in a while. But...

[00:33:07] Xavier Simon: Yeah, but it's, uh, I would even expect that it works for finance. Uh, my point is, based on our experience, uh, we were, uh, reluctant and we had some poor experience with consultants and externals. So, at the start we were, okay, let's see how it goes. And then we made clear step by step and all were matched on time, but also the support and the discussion were quietly opened.

So, that was, uh, every time, every question, uh, that we had sometimes not so nice questions, I would say, that you would ask to avoid a problem in the future, uh, were always open and, uh, with honest answers. So, that makes it easy to work.

[00:34:00] Filip Popov: Perfect. Thank you very much. It means a lot to hear that from a client. Now let's dispel with all of the serious questions. I have my less kind of difficult questions that are more related to your personal life for the audience to get to know you, a little bit of levity. Um, so having said that, what's the coolest thing in tech you are nerding about right now?

[00:34:27] Xavier Simon: Uh, frankly speaking, uh, I learned about AI quite a lot through the start of that project. So, I'll be honest, I'm quite very interested in what is going on on the AI side and what is the AI capability. So, uh, I'm fancying or understanding what that purpose or what they could fulfill. And it's very impressive. I would say in the past, we said every computer is doubling its memory every six months or doubling its capacity.

Well, with AI, the factors are quite much more impressive even.

[00:35:07] Filip Popov: Uh-huh.

[00:35:07] Xavier Simon: And, uh, moving to the second is I like physics, so when people or companies are working on quantum physics and quantum computers, if you match both, I think it will be, uh, it will be a very, very good. So, uh, that's, that's my personal look. Yeah.

I think it will not influence the industry before my retirement, but AI surely itself will influence it before my retirement. As I said, I'm expecting what, it is a worry and also it's a way of being ahead of the change. So, it may be a winner factor in our industry, but not taking into account the change that is going into AI for the moment, uh, would be something that a company would miss. So that, this is very important for me.

[00:36:03] Filip Popov: I think you're the fourth person in our episode, in our podcast that said quantum computing is something that they are quite excited about.

[00:36:13] Xavier Simon: Yes. Because it is.

[00:36:14] Filip Popov: There's a common thread there. Exactly.

[00:36:19] Xavier Simon: It is when you see the results or the things that machine could do compared to normal computer in terms of calculations, it's, uh, a lot of things will be obsolete, let's say, like this. And then if you, you had it with an AI, uh, on this, it's, I will be obsolete as a job.

[00:36:43] Filip Popov: Luckily, as you say, you feel like you're going to be retired by then.

[00:36:47] Xavier Simon: Yes.

[00:36:48] Filip Popov: You look deceptively young. That's the problem. Um, and, uh, if I called you on a Sunday at 3PM, what would I be interrupting?

[00:36:58] Xavier Simon: I would say something with my family. So, uh, my wife and I have three children, so it's, uh, weekends, uh, quite organized around the children. So, if you call this on next Sunday, it will be, uh, a birthday party, something, or, um, a scout exhibition or something. Yeah. Uh, we like to take, our weekends are mostly taking in account for the children's.

[00:37:32] Filip Popov: Of course. Makes sense. Beautiful. Thank you very much, Xavier. Thank you for joining me for the episodes of Espresso 4.0. I hope that I'll get an opportunity. Oh, I'll certainly get an opportunity to talk to you off camera. But hopefully we'll get to do this again sometime in the future. In the meantime, take care.

[00:37:53] Xavier Simon: Thank you very much. See you.

[00:37:56] Filip Popov: See you.