François Rase 0:00:05.4:

Hello, everybody. Thanks for Anaplan to inviting me here on the stage to share the story of AGC. So we made already our journey with Anaplan towards more agile planning. Of course, it's a never-ending story. I heard today, a lot of potential opportunities from Eve, on how to go the next step. So I will tell you about our story. As you can see, I'm leading the S&OP process for AGC. I'm head of the central planning team, so it's the centralized planning team where we also have local planning teams in each and every factory, that are dealing with the scheduling. The mission of our team is to provide a sourced constraint, sometimes, forecast to those factories so that they know what to produce, and when and for which customers. So a little word about the company. So AGC, it's about glass. Yes, it's François Rase 0:00:05.4: 

Hello, everybody. Thanks for Anaplan to inviting me here on the stage to share the story of AGC. So we made already our journey with Anaplan towards more agile planning. Of course, it's a never-ending story. I heard today, a lot of potential opportunities from Eve, on how to go the next step. So I will tell you about our story. As you can see, I'm leading the S&OP process for AGC. I'm head of the central planning team, so it's the centralized planning team where we also have local planning teams in each and every factory, that are dealing with the scheduling. The mission of our team is to provide a sourced constraint, sometimes, forecast to those factories so that they know what to produce, and when and for which customers. So a little word about the company. So AGC, it's about glass. Yes, it's true. Our name is as A Glass Company. So it's the initial business. But it's not only that. As you can see today, the glass is only representing a little half of our business. We are also active in chemicals, electronics and all over the world.  

François Rase 0:01:26.5: 

AGC Glass Europe is in fact the European branch of AGC, who acquired Glaverbel around 30 years ago - maybe you still remember the name - which was a Belgian-funded company that grows over Europe, and finally, yes, we joined a bigger family. When we speak about glass, there is automotive. Yes. I'm very active in [?architectural 0:01:52.7], so I choose two other examples. We are, of course, active in those famous skyscraper. A lot of glass in it, sometimes is quite special, and more typical or specialized application, like here you have a picture of the [?Giza 0:02:07.2] museum, just inaugurated in Egypt, where we use special kind of glass in a special design. So we are also active in such special projects like in the glass of your house or your everyday car. Speaking of the project where we implemented Anaplan, it was focused on architectural, a glass part and on Europe. In Europe, we have more than ten float lines.  

François Rase 0:02:33.8: 

So the float line is the name we give to the tool that is producing the basic glass. It's a long process, and the particularity of it is that it's a continuous process. So we never stop producing the glass, as soon as we light on the furnace, and this for sometimes close to 20 years. So with this characteristic, you can imagine that planning is a key challenge. If you are in a situation where you have not enough demand, how will you manage the rhythm at which you produce, what you will produce, what you will put in stock, how you will react once the demand will rise up? It's a critical point for us. We have not only basic production, basic float lines, we have several transformation steps. You can do a laminated glass, you can do a mirror. It can go into a car, into a house. You need to treat the edges of your glass. So several subsequent steps. Sometimes we also go through distribution centers, to deliver to our customers. So at the end, we have really a complex and highly constrained production network, and we have also a wide network of customers that we have to go through.  

François Rase 0:03:47.7: 

So quite a lot of challenges. That's the reason why we decided, about six years ago now, to reflect on which kind of tool do we want to achieve our objectives. There were six main elements. Complex production, I spoke about it. Also, the changeover in our production is very important. You can imagine that a furnace is a bit like the third of this room, full of melted glass, and you need to change the color. We are counting in days. So if we decide to go from one color of glass, clear glass to a green or a bronze one, we have to be sure. We have to optimize and reduce those numbers. So optimization of the production was for us a potential savings. Transport costs as well. You see what it can represent out of our same price. We have a change. The glass is heavy. It costs a lot. It protects us from the Chinese competition, but we also have to optimize our transport to our customers. While in our network, we have sometimes two or three factories able to produce the same product and deliver the same market. It can be also complex order. Sometimes we mix of products into it, so the use of not of a distribution center.  

François Rase 0:05:07.6: 

We want it also to optimize our former solution, which I will not name here today. We was heuristic and we were planning in sequence. First the float, then the transformation lines. Then we were making some changes in the sourcing, and then we were checking, okay, does this still fit with the production? It was really sequential and not at all optimized. Very important, we wanted to include financial in our S&OP process. We used to do mainly a supply/demand balance. Speaking of volumes and saying okay, we need to increase or not the production, but it was not really speaking to our top management about, okay, how will the business evolve, or will the profitability evolve, or cash flow, things like that. So it was really a lack in our tools so far, as well as the scenario planning. Basically, the previous solution was so complex to manage that it took us already more than close to two weeks to build a good enough plan, and we had one, how to imagine to make a scenario, what if analysis. While, as you know, the world is evolving much faster than in the past with a lot of volatility, and we often have to react to a sudden drop or increase of demand, also sometimes to disruption in the supply.  

François Rase 0:06:38.7: 

So scenario planning was a key objective for us. For all those good challenges, we started to look at potential vendors, Anaplan was one of them. We finally decided to go for Anaplan after a proof-of-concept phase. So we test it really, Anaplan, before to go for it. If I have to point out what were the key characteristic why we choose for Anaplan, but also why we are happy to still work with it, is about flexibility, because we didn't bought an application, at the time they were maybe not available, but we built our own solution from scratch at the time. So we went through the Anaplan way at the time, writing those user stories, but imagining a full planning solution from scratch. It's not an easy task. So of course, we had to review, we had to upgrade, we had to go through continuous improvement. This is allowed by Anaplan solution, which is your key point. Another point that I like to point out is the fact that we built the solution to support the S&OP process.  

François Rase 0:07:51.2: 

We call it the tactical model. But in parallel, when we have to make a decision of investment, in confidentiality, maybe with less detail, we had to use another solution in the past, to make a strategical model for strategic decision. Here, we build it in parallel using the same data. So we are plugged on the same data, which of course is very convenient for me as a user to not have to manage two datasets to be fully accurate, and it speed up, and it allows us also some savings. Optimization, I said it, it was one of our goal. So we do know, all at once, we have to spend a bit of time, of course, for the input data preparation to build the scenario we want to optimize. But once it's ready, we click on the button and one hour later, more or less, it depends a bit the horizon in terms of number of months, we have our results. So of course, then it's very easy to imagine that you modify one set of data, maybe the calendar of a production line, you have an alternative scenario, you push the data into the system, it takes one minute, and you click the button, one hour later you have your alternative scenario and you have built the pages which are required to compare it easily.  

François Rase 0:09:11.7: 

So all those capabilities were really powerful for us, and we could integrate in two of our solutions those financial metrics I was speaking about, cost of transport of costs, but also cost of production, or the key elements for the sales, like not exactly the margin but the priority within between the various customers. Service indicators. Inventories. We work sometimes with penalties to drive the optimizer, but all this is into the solution. The third point that I want also to highlight is that due to the fact that we went through a build from scratch solution, we went to this journey together with Bluecrux. So Mario [?Len] was part of the team at the time, and they support us in the kick-off, for sure, with their model builders. They trained our internal model builders teams. After about six months, our people were also autonomously able to really build the solution, including the optimizer part. Now, we have a center of excellence.  

François Rase 0:10:29.0: 

We have still, we are two, three years after the implementation, we have still three model builders, internally, fully able to develop further the solution, which is a key asset for us. The intuitive user interface, you know it as well. I like as well the alert-based, the fact that since we are manipulating a lot of data as input, we have also a lot of data in the output, and we need to get alerted. Where do I have to look at to validate my plan? What is my difference between the - versus the history? The difference versus another scenario? So it's always with this idea in mind of alert-based, that we have developed our solution. It was a journey, you recognized on a [unclear words 0:11:13.6] of Anaplan, but you see the one of AGC. So the idea was generated into our mind in 2020, 2021, where we also decided to go for Anaplan, together with Bluecrux. Very important was the process design. What do we want as a planning? What will be the role of the various player in this? Before to start really to model build.  

François Rase 0:11:39.8: 

In the first step, we focused on the supply planning. You may think okay, we should start by the demand planning. Maybe, yes, indeed, we would have done it if the sale department was also willing to onboard at the time. It was not the case at the time. Hopefully, it will come. So we started by that, and we could, after roughly one year of building, we could run the S&OP process with the new solution, and also the strategic model in parallel. Then thanks to this, we convinced our colleagues from the automotive to also onboard. So they build their own supply planning automotive part as well. We completed this overview with the [?S&OP 0:12:28.5] sales and operation execution. So it's therefore a tool for those local planning teams that are receiving the central plan, and then that have to execute it, but also to face all the deviation. So again, a bit alert-based control tower to ensure we are able to run the operation and execute the plan. 

François Rase 0:12:50.3: 

At the end of 2025, then we go back to the first point in the logical S&OP process, which is the demand part. You recognize the name of two application that we are evaluating as well, because we want to improve our demand planning part, our forecast generation, make use of AI, of course, and go one step further. Still, this year, we intend to upgrade our system of article strategy management. So we are a multi-echelon inventory optimization tool. The cherry on the cake, which is very important, is of course the financial planning. It's to have a full connection with the financial department. So there is already more metrics than in the past, in our solution. But here, basically, we aim at a full integration so that for every scenario that we can simulate, we have our expected operating income, free cash flow, etc. Once we will have done this fully, we will reach what we call the Integrated Business Planning, the IBP. So really evolving from the maturity of a S&OP based on volume, supply and demand balance here, to really an integrated business planning.  

François Rase 0:14:14.2: 

That's really the journey we are on. We are close to the end, but I'm sure we will have also opportunity to do even further with all the apps mentioned by Eve earlier. So all those elements that I already mentioned, on top of having better planning, it's not for the goal of having a nice planning, brought a concrete result for the company. We estimate about one or two points of EBIDTA increase. That's not that easy to estimate, because you have to imagine what would have been your planning without Anaplan. Logically, because it's much easy to observe. It's clear and we observe the improvement, and the planning cycle was drastically reduced. This is the time we need today. Once we receive an information from the same department to build a new basic scenario, and then we take the time we want to build alternative one and to derive the decision. Typically, all key decisions we are taking today in the S&OP process, like increasing the production teams on the location, rather than other one, is based on an alternative scenario versus a basic one. 

François Rase 0:15:33.7: 

So it was, of course, still a tough journey, and it was not that easy to build all this. So we had some lesson learn at the end of the journey, that I want to point out and to share with you as well. I would insist first on the process review, I explain it. We did that first together with Bluecrux. They brought their experience with the best practices from the other companies that they work with in the past. The fact of having model builders able to build by themselves quickly and to create a center of excellence is important. For sure, now there are more and more apps available, so it's probably another approach in the future than five years ago. But still for sure, there is a little bit of customization [?to raise here 0:16:26.9]. So that's important. We are using an optimizer. From what I know from Anaplan experts, here in the room, it's quite advanced. I would be pleased to discuss with another company who is also dealing with an optimizer, because it's still a challenge, because we are always pushing the limit forward.  

François Rase 0:16:47.1: 

If I have to give an advice, if you start to build this, of course, to follow the step-by-step approach. Go pacefully [sic]. Keep the control. Avoid the black box and build the complexity only gradually. Needless to say, the importance of the data quality, particularly if you are using data of other department, of course, I know my data of planning, all those kinds of the time. I'm used to them, but we need to use the data of sales department, of manufacturing, cost of it. Basically, the really important point is to build the trust, via the data quality. The other people recognizing their data into your tool, I speak of the other department, they will have trust in the results of your plan, particularly if you are using an optimizer that can also look like a black box from their point of view. So very important. Maybe a detail here is the online user documentation. As a user, I'm not making model building, but on every page that I'm looking at, that I'm using, I can write the necessary documentation for that another user in my team is also able to use it, and if we do a continuous improvement, I do it.  

François Rase 0:18:05.8: 

So today, it's not the model builder which are building the documentation, and particularly not on an off-site document, that you never go back to it, and when you go back, you realize it's not appropriate. We put it directly into the page. Last but not least, I think, at least in our case, that having the partnership with Bluecrux was a key factor for the success of our implementation. Yes, that's really what I wanted to share with you. I know that we were a bit squeezed by the time, so if you have questions, I will be pleased to discuss at the coffee break. I hope you have a good view on our journey.