Webinar: How AI Out-of-Shelf Recognition Can Improve Retail Efficiency

Good morning, everybody. Uh first of all, I would like to thank you for attending our webinar, which is dedicated towards how an AI image recognition can help you to enhance retail efficiency. And today, during this webinar, I am joined with Victor Hart, who is our senior account executive, who is responsible for our business development among North American markets. And myself, my name is Andy. I'm a sales team lead at Liftopia AI. Before we jump to the main topic and uh key agenda, I want to just to give you a very brief overview about Liftopia AI and what are our credentials to discuss on the matter. So, uh first of all, we've been on the market for the last 15 years. Over this time, we have implemented more than 200 projects across 30 countries. And currently, our operations are supported by 120 employees worldwide.

As a matter of fact, to give you an idea in the nutshell what we do, we provide one unified platform that helps to automate key processes within the supply chain. Our platform consists of five standalone modules, which can work, of course, on the standalone basis, but also complement each other and supports various complicated supply chain workflows within the retail. So, the key parts of our platform are assortment performance, a module which is designed to help you create your assortment strategy and decide what kind of products should you be selling and where, and what kind of products you shouldn't. Shelf efficiency is a core of the today discussion, and this is a system that helps us to maximize the use of space on our locations, and of course, maximize profitability on the space that we have. Inventory optimization is one of our core modules, which helps to fully automate replenishment process and eradicate your

lost sales and minimize any potential overstock that you might be having. Promotion intelligence helps you to have a better stock management during your promotion cycles with the help of AI. And last but not least is a customer loyalty, one of our modules, which by itself uh not only helps to execute various mechanics of the loyalty on a fully automated basis, but also helps you to aggregate a lot of information about your customer, learn who is your ideal customer, and how to communicate with them. So, again, establishing a direct communication channel with your customers in order to be able to also influence your demand and be very relevant to the customers you're interacting with. And uh with regards to our customers, we work with a wide array of customers from a wide range of industries. So, of course, we have a number of customers from the grocery and a number of global brands such as Spar, Carrefour, Smart & Final, but of course, we have a customers from a different sector as

well. So, we work with DIY customers, we work with a number of uh customers in gas and oil confectioneries, and pharma, health and beauty as well. So, without uh further ado, I would like to pass a floor to my colleague Victor, who will speak a bit more precisely with regards to the solution that we are going to be demonstrating today, and of course, retail challenges that can be tackled with it. Victor. Thank you very much, Andy. Um so, I'll take over the screen control now, and I'll dive into some of the core functionality that we're going to be looking at today as it relates to how AI can help you um to streamline your processes and improve the underlying profitability of the retail business. So, what we do from a shelf efficiency solution, which is the module that we're going to be taking a small look at today, the core functions are macro spacing

management. So, like Andy said, how do we make the best use of our space that's allocated within each one of our retail locations. Could be very large supermarket, hypermarket, that type of location, could be a small convenience store, or anywhere in between. So, how do we manage the store layout? And then how do we place products into the store in the optimal in the optimal way? And then of course what we're focused on today is leveraging AI and machine learning to say, "Okay, we know what product we want to place where, how do we ensure that we're adhering to to that strategy, to that logic?" And then again, the core goal of the solution is to improve revenues, improve profits and things like that. So, from an AI out of shelf standpoint, what

does that mean? Well, how can we enhance our efficiency? How can we reduce the manual processes that are needed to ensure that we have the right product at the right place within our shelf? How do we prevent lost or potential lost sales? You know, if we don't have product on the shelf, then obviously we can't sell that. And so, how do we identify that and make sure that we get the right product on there? And then compliance, not just having product on the shelf, but having the right product on the shelf that adheres to the strategy that was um uh dictated by the central merchandising team. So, that we if I'm a central merchandiser and I'm designing a number of planograms, how do I ensure that the employees within the store are actually placing the right product on the shelf that adheres to that strategy. And then

of course, a better customer experience. Um, if I as a customer go into a store that consistently has a lot of empty spaces on the shelf, that's not going to be the best customer experience. Um, and I might uh look to a competitor or something like that. So, how do we keep that customer experience as good as possible so they keep coming back and um are repeat customers within our stores. So, what we'll look at today is how can we leverage AI to improve shelf availability, ensure we have accurate planogram compliance, um and then that feeds into inventory management as well. So, if for example, we identify that we do have some product missing from the shelves, is it an issue with inventory um which our other solutions can help with, or is it simply an issue with execution at the store level? So, without further ado, I will jump

over into the Leafio platform and will walk through some of the aspects of the shelf efficiency solution specifically related to um to AI out-of-shelf recognition. There's going to be a lot that we can't get to in today's webinar. And so, of course, um if you're interested in some of these other topics, please reach out to us. We'd be very happy to set up um individual meetings to discuss your, uh, unique business needs, um, but I'm going to be focused on, uh, the AI out-of-shelf recognition today. And so, from that standpoint, the first thing, of course, that we have to do is going to be building out a planogram. So, um, I've got a specific floor plan opened up here. I'm going to open up one, um, specific planogram. And so, in this example, and again, we have lots of other, um, types of businesses, um, from DIY, grocery, um, pretty much

anything in the retail space we've worked with over the past 15 years, uh, but this specific example is a, um, kind of convenience store where we are selling some, uh, uh, in this example, some juices. As my system is loading, um, it's early in the morning here in the US, and so maybe my laptop is, um, trying to take its sweet time. So, we'll give it just a moment. When I build out the planogram here, so one thing that I want everyone on the, uh, the webinar here to keep in mind is there are a number of different strategies that we can take to build out our planograms from a purely manual approach to a fully automated approach across, um, lots of different, uh, locations. Um, so here's an example of a planogram, and I'm going to pull these products off of there. And so, I,

as the central merchandiser, can assign rules to the system to take data from the client's existing systems, uh, typically an ERP system or something like that, and, uh, then automatically build out the planograms for me. And this could be a planogram that's assigned to one location. It can be assigned to 100 locations. Um just depends on the strategy of that um that company. And then very simply, I can leverage the rules that I've already given the system and have it automatically build out the optimal number of facings for me, which can then be sent down to the store location um for execution. And it's taking a second um looking at all of the data that's in the system, the sales history, the the assortment matrix, things of that nature. Um and then it's going to build out the

optimal number of facings for me at each store that I have this planogram deployed to. Um I can save these. I can work on them. I took an automated approach in this example. I can manually adjust it if I want to, if I want to reduce the number of facings, something like that. Um very easy for me to do so. I can save it and continue to work on it over time. And then once I publish it, that's going to then send this down to the um in-store employees for execution. Um and we can do that several different ways, but probably the most uh the optimal way is going to be via um our mobile application, which is functional on either iOS or Android cell phone, laptop, um whatever type of device is available to you. And so, give me just one moment and I will sync my phone to my laptop here

application. And what it's going to do is assign um assign tasks to the in-store employees. So, if I pull up a specific store, I'll then see if I have any tasks assigned to me to either build out new planograms or update existing ones. And so, I could look at um any one of these. I've got some new tasks that we can see here. I've got some in progress, um one that's expired, and we can monitor and um and um make sure that those tasks are being executed both at the local level as well as in the mobile application. I can also um look up just a specific planogram. Oops. I can search all of these planograms that have um tasks assigned to me and pull up a specific one. So, here we can see how the planogram was designed

by the central merchandiser in the web application, and I can see that here within my mobile app. So, I can um tell what specifically I need to adjust. If we have a very large store layout, for example, I can um look up a map of that store layout and see where specifically within the store I need to make any adjustments. You can see this green icon here. That's going to be the piece of equipment that I need to make some adjustments to. I can go back to that. Um see the uh the products and how they need to be laid out, and then we can link into the uh the um camera on the phone as well. And then I can see I'll take a photo after I've made my adjustments to the uh planogram, and once I submit that, that then gets sent back to the uh the web app so that the central merchandising team can confirm

that the necessary adjustments were made. So, this is where the AI component of the application comes into play. Um, rather than just having to have, you know, Andy as the central merchandiser reviewing all of the photos that I, Victor, took at the store level, we can leverage the system to automate that and simply identify issues with it. Are there empty spaces? Are there, um, uh, products that were placed on the on the realogram that don't match the planogram that I made, um, or that Andy made as the central merchandiser. And there are several different ways that we might be alerted to that. So, one would be through, um, our, uh, our business intelligence solution, which I got signed out of there. So, let me just pull that back up. Give me one moment here. So, um, there's a full business

intelligence solution built into, uh, the platform that can report on a wide variety of of different types of key metrics that, um, that you might want to, uh, monitor over time. The one that's most relevant for us today is going to be the AI out-of-shelf recognition. And so, what the system will show us is if we have any out-of-shelf space detected. So, as the central merchandiser, I can monitor this, um, come back to this report over time. Maybe we've got some kind of recurring cadence where we update our planograms on a monthly basis or, um, every couple weeks or something along those lines. And then we can see if we have any out of uh shelf space detected here. So, um we've got 138. I might want to dig into that a little bit

deeper to figure out what's going on there, and that's where our table comes into play down here at the bottom. And so, we can adjust how this report is displayed. Um so, I might want to be able to see the planograms themselves. So, for example, the way that this is set up right here, I've got my various stores that I'm monitoring. I can open up these stores and see the specific planograms that have any out of shelf um space detected on them. I could add in the date as an example. And so, maybe I want to see, okay, I've got this store. Open this up again. Um and what date we detected the out of shelf um space on this planogram or this planogram or etc. And so, we can leverage the BI solution to um monitor kind of the overall aggregate

um status of all of our planograms, and if we have out of shelf detected across the entire chain, I could filter down to specific stores, I could filter down to specific planograms, um specific employees who are responsible for filling in and stocking those shelves if I wanted to. Um and then I can, of course, bookmark those as well. So, if I build out lots of filters in this reporting, then I can always um go back to that very easily and find it very easily. So, the BI tool allows us to monitor that um at a very high level. And then, if I want to look at it from a store level, I'll come back to this store layout that we have here, and I'll just adjust my view. Um I can have the system show me any pieces of equipment that have out of shelf space identified here. So, any of

these pieces of equipment that are in red, the system has identified that we're missing space, which of course gaps in the shelf is going to be potential loss sales for us. And so, if I open this one up specifically, this is the exact piece of equipment that we looked at earlier in the mobile app. Um if I go into the details on this equipment, I can see the uh photo that was taken in the store and sent back to the web application. So, here is just the the core photo. Um if I click on this button, then we can see that out of shelf recognition. So, the system is identifying that we built out a planogram centrally um in the web application here, and then based on the photo of the realogram that was sent in from the store, we've got some gaps. So, there's missing space here, missing space here, here, here,

etc. All of that missing space is going to be uh potentially lost sales for us, and then we can dig in a little bit deeper to try to understand what happened. So, was it a inventory issue? Was it an execution issue? So, for example, this item here at the top, uh the central merchandiser uh designed it to have two facings. There's only one facing that we can see in the realogram in the store. Was it an inventory issue? Uh when I look at the the details on this product, we've got a stock balance of 50 units for this product. So, it wasn't a inventory issue, more likely an execution issue. So, that's the key is to give us this kind of insight and leverage artificial intelligence so that we can identify these issues on an ongoing

basis and ensure that we're keeping our real grams in the store as closely aligned with the optimal number facings and the optimal planograms that we've designed in the the web application here from the central merchandising team. We looked at the the photos via the mobile application. If you have cameras installed into your stores as well, we can integrate into that, which would be more of an ongoing automated process. And so, via the mobile application, we can anytime we send the task down to the store level and they take a photo, then we can report on that and and monitor it just like we're looking at here. However, if you have a the the cameras set up throughout the aisles, then we can schedule recurring calls to those cameras to take photos, send them back

to the web application, and have that as more of a recurring process. And then, of course, the whole goal of this is to ensure that we are increasing our revenues, reducing the lost sales, improving sales, things of that nature. So, just jumping back into the uh business intelligence tool here in our like-for-like analysis. Um over here on the right, we can see our sales or sales per meter or however you want to um analyze those type of numbers, but we want to leverage this technology so that we can ensure that we see improvement like this over time. We can see our sales numbers increasing over time, our profits increasing over time, etc. because that's going to be the whole goal of the solution is driving improvements in sales, improvements in sales per meter, um

reduction in our lost sales, etc. So, that was a quick look into the AI out-of-shelf recognition that Leaf can help um empower. Um I'd like to open the floor if there are any questions that can be submitted. We'd be very happy to answer those or if you'd like to um schedule some time to take a deeper look into your specific business, your specific use cases, we'd be very happy to schedule some time with you as well. And at the same time, I would like to mention to everyone that we you can reach out to us at our website www.leafio.ai. And of course, you can schedule a meeting with a dedicated professional on the topic of your choice. So again, Leafio as a company, we help to solve a number of issues within the supply

chain. So, merchandising is indeed one of the pressing issues and we get a lot of interest from numerous verticals of the retail business, but of course, there are other core problems which we help to solve within the area of replenishment, within the area of choice of assortment, within the area of customer loyalty and promotions. So, again, if you have any questions, any queries, again, feel free to write us to sales@leafia.ai, fill in the request form at our website, or even contact through a phone. We will be happy to discuss and learn about your issues in more detail. So, I'd like to thank everyone for participating in the webinar today and of course for your time and really hopeful to meet you all again, maybe first in the virtual environment, but of course in a long run, hopefully also in real life. Many, many thanks for your participation.