Espresso 4.0 by
Wizata
Thanks for joining me today for an Espresso 4.0 episode, Floridian. Super happy to have you. We were super glad to meet you at Dah Tour a couple of weeks back. And thank you for accepting my invitation.
Why don't we start by having you tell us a little bit about yourself and your background so that the audience gets to know you, and then we can get right into it?
About our Guest
Yes, sure. Thanks, Philip. So thank you for the invite. It was a pleasure to meet you a few weeks ago. I am Florian, I'm a consultant. I have a background in the supply chain.
I work about ten years in the industry for different kinds of sectors. Automotive, railway, retail, and manufacturing. And then, a few years ago, I decided to move from corporate supply chain and move to consulting.
And since then, I've worked in many areas in Europe, projects in Mexico, for instance. Maybe we will speak about it. And as well, I am also a teacher at university for eight years now, teaching warehouse management in France, and I had cooperation in Paris about innovation in the supply chain.
Glad to have a scholar, a gentleman, and a scholar. So let's start with a high-level subject, and then we can zoom in and dissect it later. You've worked in supply chain management for projects for the better part of a decade right now.
How can AI Help Optimize the Supply Chain?
And we keep hearing that artificial intelligence (AI) can help us optimize supply chain management. Can you tell us how we can do that? How can we improve on the things that we're doing right now, specifically and especially if you have some examples that you can tell us?
So, first, supply chain. I would say this concept was revealed in the 90s or the beginning of the 2000s. And then, when we connect transport to logistics, we start to speak about a supply chain.
So, supply chain, in the end, what we intend to do is to bring raw material to a product and to deliver it to a client or a customer.
Here is something really basic. Sometimes we just say, okay, when I speak about supply chain, it's just taking a parcel, a palette, and delivering it. But many people don't know what's happened behind the scenes.
Three Key elements of the Supply Chain
And here is where the supply chain is really interesting to me. Basically, I would say that there are three key elements I would like to speak about today to have an efficient supply chain.
First, if we deliver the product, we must ensure the integrity of the goods we want to bring to the client. Because after all the effort we will make with different chains and processes if the product is destroyed or damaged, it will be really bad for the client.
So here, first will be the integrity of the product. The second element is the lead time. It's not to deliver as fast as possible. But it is to deliver a product like goods in a time frame that a client is willing to pay for or is willing to accept.
So we have seen now with Ecommerce, you can order by twelve, and in the next 4 hours, you already have the product at home delivered by a postman.
So that can be great. But then you will also speak about a sustainable supply chain, and we will see how to combine the factors of delivering fast and optimizing the volume to be sustainable.
And the last one, the third is cost. So you need to be cost-efficient, and for that, we speak about volume optimization as I said, the challenge will be how to optimize, reduce the footprint, reduce the cost, and simultaneously deliver in a really short period of time.
Why is Stock the Necessary Evil?
So here are the three main concepts, and it reminds me of something I've been working on. When I was a student, I applied for a master's degree in supply chain. The process was first to send a CV, application, motivation letter, and why you would like to do so, and then we had to write a thesis.
And my thesis, the topic I've chosen was why the stock is a necessary evil for the world. And here we can see that, okay, AI is definitely a lever to optimize the supply chain. As you know, you need to have stock somewhere to allow your client or customer to get the goods in time.
So, in the end, we can introduce the concept of AI that we need to optimize the supply chain flow to reduce the cost and the waste of material. I would also like to speak about AI and how we can use it in the supply chain.
So I think the next question would be about what predictive methods or what we can predict with the AI when we master the data.
Using AI for Predictive Methods
I would say that AI is a way to use data. We have a new goal for this century, and predicting what could happen next is a way to optimize the supply chain. Because as I said at the beginning, we need to optimize the cost, and the lead time, and at the end, we need to bring the product in good shape to a client.
You cannot process any AI if you don't have any data. I think it's a basic, the main ingredient for the recipe. Let's go to the recipe and see how we can make something good with that.
So, when we speak about data, and I think in your experience, if you want to apply any algorithm, any AI, to get the best value of the data, you will need first to clean the data; you need to see where is the deviation, what is scrap, what is not in a good variation.
So a lot of noise. Exactly. We need to reduce the noise. And this will be the first step if we go to an AI or data project. I would also say that we can have different steps for the data.
Creating the Data
The first step is to create the data. So, in some cases, you are trying to optimize the supply chain process, like picking the product or preparing an order for your client. So, what key levers do we have here?
The first one is for the picker, the worker who will prepare the order to optimize the movement in the warehouse. And here, we can have the first method, which is to collect the data based on the path established by the WMS - the system which controls the stock and the ordering process, the Warehouse Management System.
We're using the data we had in the system to optimize the path, optimize the movements, and get productivity.
Connecting the Elements
The next topic is okay, how we could connect the worker and the other element of the warehouse. You have colleagues, a forklift, a truck driver, and different departments working in the same warehouse. And there we started to think about, okay, how to collect the data that, for now, we don't have.
And we started to integrate into the surfactant processes any kind of internet of things. To bring life to an asset, which is just collecting the data for now. But you cannot explore what's behind it.
I would say AI will also bring the data to the next level. And if we can make a correlation between existing data and a new one you create, you bring value to the supply chain. So we need to create data, then we need to process or mine the data to see how we can get potential, measure it, and then evaluate the results.
Getting Data History
And the last one would be, I would say, how to get a history of your data to see what has happened in the past.
And then, here we can introduce another concept which is machine learning. So you want to use the past events not to reproduce a mistake and maybe to see your habits in a warehouse, in a transport chain, in any place, just to optimize what you do and do better and better.
Thank you. If I can summarize and distill what you just told us, and you've told us quite a lot. Thank you for your extensive explanation. It is somewhat of a recent development that we're thinking of as a supply chain in the context of more than just transport and logistics, but as many chains from the raw material producer all the way to the end. customer.
And the way that the AI optimization helps is with the data between all these so far siloed categories, there's a data flow traceability, and when applying AI and machine learning, we can find correlations, find trends, and have information at the unprecedented speed to cut the lead time, make sure that the product gets to a client as fast as possible and in good conditions.
And, of course, cut the costs in the process by optimizing everything. And if we achieve all of that and get some sustainable processes, I think that’s the supply we should have tomorrow.
Having an Insight into the Production Process
Now, a view of production is a key element in this supply chain, right? And having a good insight into your production. The smart monitoring solution is paramount but not enough in and of itself.
Where we're heading right now is having certain predictive analytics and predicting how your production will act in the future. And one of the key value drivers there is predicting maintenance, avoiding downtime, and so on and so forth.
You, in fact, worked on a predictive maintenance project, and I wanted to, without divulging any names, if you will, you can give us a little bit of an insight into how that went and how it affected the value chain.
Thank you, Philip, for the question. Yes, actually, I worked on a few predictive maintenance projects. I would like to pick up just one. it was for a national railway infrastructure company in Europe.
The aim was to maintain the rail, the catenary, and all the systems to bring electricity to a train and the signalization. So how to smooth the traffic with the right signalization and to be on time.
So here's the target of my client. He already had all the systems to create and collect data, so he had a system for the rail, a system for the civilization, another for electricity, and a power system for the catenaries.
All the ingredients were there. He just needed the baker. But more than that, actually, he needed to find a way to make a correlation between all of these individual systems. And here started the project.
Data Analyzing Platform
The project was to have a platform where you can analyze the data, the lifecycle of your assets, and your network and see if we can somehow connect a failure type from one system to another and how.
When the signaling station system is inefficient enough, we can see that the rail can be used normally, so we have tried to improve this correlation with this platform. And to improve the state for the client quite a bit as he will be managing, I would say, millions of Europe assets in the country.
And for sure, I'm just speaking about the network here. But then you need to think about the client and the network user.
If your library network is not optimized or working properly, we all know, unfortunately, we all have these adventures for Christmas time. We needed to take the train to come back to our family, and then the train got delayed.
Or you have this nice voice on the mic telling you we had a technical issue on the railway and need to postpone the departure. So here is what, in fact, is the biggest take for the project.
So predictive maintenance is one thing, but the big picture is to improve the network and, in the end, give the customer good service. To just take the train from A to B, being on time, and yes, that was a target. To be on time for Christmas with family.
In fact, I can, speaking from personal experience at least. It doesn't matter if it's Christmas or if it's just a regular Tuesday. I can tell you I hate when I hear that announcement.
There's a maintenance problem and a delay as a consequence. So on behalf of every public transport user, thank you. I think this is in particular where AI in predictive maintenance shines in an example of a process, let's put it that way.
Not necessarily a production process but a process with many nodes, tags, and data points that you need to correlate and make very fast calculations. Also, predict how they will behave depending on if this data point way over there in the beginning of the process or the middle of the process, how that goes one way, or as anomalous as that would affect everyone else, right?
Unlocking New Streams of Revenue
Some industrials are finding out that by tapping into their data, they can unlock other streams of revenue. Can you give us an example of one of the projects you worked on where a customer or client of yours has discovered this?
Because we spoke about data and how it's an important recipe, we've mentioned this a couple of times. But apart from optimizing your current business, can it also bring value in terms of discovering other ways to produce revenue?
Yes. Thank you, Philippe, for asking the question. Because basically, when we speak about supply chain, when speaking about maintenance, we focus on the cost, and we don't focus on the value of the revenue.
And here, I would like to be a bit more optimistic because before being a consultant, I worked for different corporates and had a chance to work for the logistics provider within the innovation team.
Technological and Bussines Innovation
And there we had basically two approaches. The first one was so-called business innovation. It was kind of the good part of the department where the target was to bring new ways to get revenues with new ways of pricing and changing the company's business model.
So they were, let's say, pulling for more revenue. And the other part of the team, where I was, was focused on technological innovation and bringing the technology into our logistics and transport processes.
When you speak about innovation in logistic processes, you need to invest. You need to buy the technology or rent it. You need to do a lot of research to see how you can optimize the lead time, product quality, or any other parameters.
And here, it was a bit frustrating for me that for my other colleagues that were focused on the revenues, it was quite easy to get any budget for marketing, advertising, and delivering a new offer on the market.
And on our side, we go back to the CEO every year. We have this project. We need more money. We are a bit short because it's research and development. It's not like a regular project. The ROI is unclear.
We are building our own ROI. If you want to queue innovation, you need first to focus on the ideas, try them, and make proof of concept, and then when you find out the concept, you can speak about return on investment.
Here, I was a bit frustrated and tried to combine technological innovation with business innovation.
The Application
We were working on something quite cool. We were trying to make it with image processing. We wanted to use this kind of algorithm to count the pallet. The business case could be in a warehouse or on the track.
So the idea, first the business case we had, was to secure the freight in the trailer, in the lorry. And for that, we used a Kinect system. You might know, used by Microsoft for this game, where the system can catch your movement, and then you play with no remote.
We had to invest a lot in R and D, and we worked with a lab to develop the system and to get the right level of measurements of a pallet because basically, we were trying to measure how many cubic meters were in the pallet, how many pallets we had in the truck.
Then, check when the truck driver goes to a shop or a warehouse if what he had to deliver has been properly delivered, if there is no missing pallet in the truck, or just to check if everything is okay.
Then we tried to see, okay, with this system, how we could generate revenue. And at the same time, in this company, we were also working on building a marketplace for transport. And then we had, I would say, a crazy idea to connect the marketplace with the device.
The Business Case
The business case was quite easy as it transports at a low margin, basically. You need all the time to fill your trailer as much as possible. And sometimes, you can take a full trailer of 33 pallets and deliver them to your client.
And sometimes, you will have some batches of 5,6,10 pallets and make like the milk run, but you will drop the pallets. And there, we wanted to have a smart trailer that would be able to speak directly with the market, trade market, and freight exchange, basically.
It needed to say, okay, I'm a trailer, I have ten cubic meters empty. And just tell me, if you localize me on the map if somewhere I could take some other pallets, buy them and then deliver them to another client.
So we've tried to generate revenue with this system, and thanks to that, we could sell the idea internally to my Transport Department colleagues to agree to test the material. So this was one idea or one of my experiences where we tried to combine technological innovation and make revenue with new technology.
And here, I think we could dive deep into different kinds of industry sectors to find a way to make more revenue with the data we collect and have. That's a pretty cool project. Basically, you're optimizing the real estate within the freight truck.
Yes. That's pretty cool. Okay, Floridian. Thank you very much!
