 Good afternoon all, what? Thank you. More than religion, more than politics, more than sports, what's that one question that unites us all? It's a very simple question. Where the hell is my package, right? We have all always waited for one or the other delivery packages for us. Logistics is one of the most underrated issues in our modern world. If you think about it, the number of people who are hungry every day in the world multiplied by the amount of food they need to take every day is less than the amount of food wasted every day, which basically means world hunger is a distribution issue, not a production issue. So why is logistics that complicated, right? Let's have a look through the journey of a package from any point A in the world to any point B, right? What happens at each stage? Let's just dive open each point of the package, right? So let's say when you place an order, some of the issues that the logistics companies face at that point are that they do not know the size of the package that you are placing an order because that information is only with you. They don't know when they should be picking it up because you may not be at home, you may be available, do you want a convenient pickup time? And that adds just to a lot of variability, right? Do you know, by the way, who makes it so complex? It's all of us. 20 years back, all of us, by that I mean the consumers were never part of the supply chain story. You used to drop a package to some FedEx port and you used to buy something at a retail store. Today, you wanted to be picked from your home and office and your goods to be delivered back to your home and office. This creates a level of complexity absolutely unseen before in logistics, right? So consumer becomes a part of the supply chain and all of these problems comes just at the beginning. Then when it goes to a warehouse, all the packages from all of your homes are collected together, they go to a warehouse at the end and at that warehouse, now you have another set of issues, right? What are they? First of all, how should these all group of packages should be shipped across air, freight, surface, whether it should go on a bike, on a truck, whether it should go today, how should you manage taxation? And this becomes an immensely complex problem. Second, now a typical company will have a few hundred distribution centers. What's happening at each distribution center? Now, each distribution center does not work in isolation. It's exactly like a stock market. The performance of one's thing in one sector will affect the stock price across the sector, right? So can the movement at any one point of distribution completely change the flow at another point? And how do you know that when you don't even have that kind of visibility? And then, even if you resolve that, every single minute that a package is delayed, it's not just the transportation cost that's drastically increasing, it's the storage cost. Storage contributes to between 40 to 60% of the cost that is required at any point in any logistics movement, right? So every time you delay shipping out a package from a warehouse, you are effectively increasing its storage cost. Now, let's say package made through all this journey, fought through the complexities at the first mile at the warehouse, now comes the most difficult part of the logistics, the last mile. That is, that is the package actually reaching the intended destination. It has several issues. Now, of course, the first issue that we have always faced is the timely delivery of our packages, right? Now, if I get an SMS in the morning saying, hey, my package is out for delivery, that's not really useful. Do I go to my office? Do I not go to my office? Do I go pick up my breakfast when I got to be at home? When the package comes, what do I do about it? Second, the dispute in businesses that, hey, has the package been delivered? If it has been delivered, has it been delivered in good condition? And when, if you want to see how complex this issue is, you should really go to a warehouse of a third party logistics company. You'll be surprised to know that as much as storage they have for packages, they have storage for a bunch of papers reconciling the proof of deliveries across the supply chain. And then finally, you have, can we deliver the package in the first attempt? Because when you make multiple attempts to deliver a package, it's not just the inefficiency. There is a social cost that comes in. Is, think about it, is that $3 connector that you ordered from Amazon worth making three attempts to deliver at your home, the social cost of it, the pollution cost of it, the carbon footprint of it. So it's not just about the efficiency, it's about being responsible. You want to do it right the first time. So these are the complexities that, you know, that logistics faces at each stage, the first mile, the middle mile, and then the last mile. But then the question is, are we here to just describe the problems? Absolutely not. We're here to talk about how can we bring a bit of magic in motion in the entire logistics piece. Now, what do I mean by magic in motion? Well, think about it this way. Can all of you tell me what is this? Of course, you can tell me it's in hand. Can you take out your phone and tell me by reverse, doing a reverse video search? Absolutely not. Because figuring out what is there on screen is a problem of visual cognition that is suited for humans. But if you think about it, your supply chain, as we just described, is a bunch of fast moving variables. It's nothing but a multi objective, multi variable meta heriostake optimization problem, exactly like a stock market. It was designed to be algorithmized. And with the recent advent in data technology, the map data sets available to us, we can algorithmize this entire decision making. So the question is, can we automate every single piece of decision making that is required to transport a package or a person from any point A to any point B? No human decision making involved in it. And if you do that, let's look at how the sum of the advantages we gain at each of these steps. Let's go back to our first mile, right, where we talked about the pickups, right? Now let's see if you can do. So if you create an IoT device, which instead of scanning just a barcode, also measures the length, breadth height right at the point of shipment. Now you have introduced that digital information, which makes the whole planning piece effective. And towards this, Locus has built its own handled hardware that instead of just scanning a barcode, also measures the length, breadth height. So you start digitizing at the very start. Second is your dynamic planning. Instead of instead of telling the consumer, hey, I'm going to come in the second hour to pick your order when the consumer may not or may not be there, ask the consumer, when should I come for a pickup? And your system should be then smart enough to factor that dynamic planning. Can we give like, you know, as precise information as, hey, I'm going to pick it up in this one hour slot that you select. And that drastically, you know, reduces the friction required in just the order collection before it even reaches a warehouse. Now let's see what happens at a warehouse. At a warehouse. What you can do is that you can run a bunch of simulations if your entire digital network that, hey, during the festival season, just around the cherry blossom time, when I have a bunch of people coming into the airport, what kind of flight network should I run? And this should not be just based on experience, but in 2000 should be based on data. You can run a sensitivity analysis. You can run a bunch of simulations. And this is today possible because of the advancement in data, machine learning, optimization techniques, and as of last year, deep learning methods, which allow you to learn a lot from unstructured data and generalize it. Second, there are, thank you. That's good logistics. So second, if you had each point in your supply chain, you have the relevant hardware that can that can tell you what is the visibility. Because let's remember, yes, we want to answer the question where the truck should be. But to answer the question where the truck should be, it's very important to know where the truck is, right? So the first step to automated decision making is visibility and transparency. And that gets bring in with a large IoT and sensor networks. Previously, you could have been collecting this data, but we just did not have methods to process this into meaningful and useful insights. Third, cost modeling and optimization, right? Countries have double digit percentage cost of their entire GDP as a piece of logistics. A small, a few basis point movement in the cost of logistics as a percentage of your GDP can change the fortunes of the company. As I mentioned, there is a first principle reason of this problem to be solved by computers. Because as I said, all of you can recognize what is this, but can you tell me what is two to the power 98? No, while a $3 computer can. Logistics is a problem designed to be solved by computers. If you feed the right amount of data in it, have the most amount of propriety algorithms, which constantly adapt to your business, you can you can bring cost optimization and modeling to a double digit percentages. We work with public market enterprises throughout the world who have been running logistics businesses for a decade based on experience. And simply because of the newer algorithms, they have been able to reduce their logistics cost almost between 15 to 23 percent. And these are large public market companies. It's not that we are doing some magic. We believe we are, but it's simply this problem was always designed to be solved by computers. And now we have just dedicated an exceptional unit to 20 percent of my entire team is PhDs from various schools like Carnegie Mellon, Columbia, Ohio. The founders were building the AWS machine learning network before this. So we have built large scale public machine learning systems along with having propriety algorithms, attacking a problem which was always designed to be solved by computers using the most recent available data set. These three combination just makes it an amazing problem to be solved by computers. And now finally at the last night, this is when your package is really available to you. Imagine a world where your courier company, instead of sending you an SMS, which says, Hey, I'm going to deliver your package today. It says, Hey, I'm going to deliver your package between four to six p.m. And at four 15 p.m. sends you another SMS with a tracking link where you can see your package coming live to you for the last half an hour reduces anxiety. You are available at the last point to actually receive your package removes the reduces the turnaround time for the company as well. Right. A simple signature for you, you know, on their tablet or via an OTP or via a simple picture makes the whole reconciliation very simple. It's not just electronic proof at that point. It also earlier you digitize the data in the supply chain lesser chances of it ever going into fraud. The third most optimal the third most important point now because today let's say this group of people ordered a package from Amazon tomorrow this group of people ordered groceries from OSIX in Japan. Now because consumers are split at different different points in a city and different consumers order different days. The traditional last 200 years of methods of dispatching logistics just don't work. Now you cannot have static routes going from each dispatch center based on everyday's demand. OSIX needs to calculate how many trucks to run to get your deliveries to your home every single day probably multiple times a day and that requires sophisticated routing engines but that gives us two things one it gives us efficient routing that means you're spending less to deliver your packages and second you're ensuring that your first attempt rate of deliveries are increased which effectively reduces your carbon footprint making us all a considerably responsible citizen. It is our firm belief that in very near future by 2025 there will be large companies whose chief supply chain officer will not be a human but will be a software. On that note as is our business we always deliver on time. Thank you.