Loadsmart CEO Ricardo Salgado attended the Trans-Pacific Maritime Conference hosted by JOC. Check out the interview from the event and hear the origins of Loadsmart and the fundamentals of the algorithms that power the automated pricing, booking, and shipment of freight.
Here’s the transcript:
(00:20) AB: I’m Alessandra Barrett, Senior Content Editor for JOC and I’m in our 2019 TPM Conference with Ricardo Salgado, CEO of Loadsmart. Great to have you here.
(00:30) RS: Thank you for having me.
(00:31) AB: With the recent spot market fluctuations, everyone is probably looking for something a little more predictable. What inspired you to create Loadsmart?
(00:43) RS: I actually started very young. My parents owned a paper company and during my internships there I would go out there and realize that in this industry, how we moved goods was highly inefficient. We were in the middle of this technology revolution and as an engineer, I looked mathematically from an engineering perspective and a process perspective with the latest technology, I realized there has to be a better, more efficient and socially responsible way to move goods from point A to point B.
(01:17) RS: As I started going to college and work, I realized there are massive investments in data infrastructure, for example 5G, but the truck drivers today, instead of 25% having a smartphone, now 95% of them have a smartphone. So as you started seeing more connectivity and you can hold them accountable and part of this sharing economy and dual rating systems, I realized there’s definitely an opportunity where you can put it all together and all of a sudden it’s an industry that’s huge, fragmented, lacked technology… and it all just came together at the right time.
(01:55) AB: Loadsmart uses AI for truck brokerage. Can you walk me through how the AI is applied?
(02:03) RS: Absolutely. Our core technology does two things. Number one is we provide instant pricing and booking capabilities nationwide. So at any point in time, just like you book a flight on Expedia, you can say, “I need a truckload from New York to Los Angeles,” click a button and get a price immediately. Now, New York to Los Angeles is a highly trafficked route because they are metropolitan areas. But if I go South Dakota to Maine, there’s not a lot of volume there. To generate that rate, we estimate it programmatically based on roughly 385 different factors. Our AI and data science learns every single time we render a price and iterates through it. Think about each one of these features as a different knob. If we render a price and it’s accepted or not, then it gives us feedback and our learner models kick in and start adjusting. It asks, “what if we had shifted this one or the other one? Would it be more accurate versus where we actually sourced that truckload from our partner network.” So, that’s one element.
(03:09) AB: Oh, I see.
(03:09) RS: The other element is identifying — it’s our sourcing algorithms. Our sourcing algorithms, which is on our supply side, is identifying the ideal carrier or trucking partner within our network or outside of our network. It’s asking itself which is the best carrier at that point in time. Which one is moving in the right direction with the right driver for the right rate in that same time, etc. So those sourcing algorithms also start learning from that carrier base, powered by AI and data science, asking “Did this one work? Did it render correctly? Was it accepted?” It starts learning, and ultimately it starts optimizing both sides of where we sit in the middle.
(03:50) AB: Well, thanks for taking time to sit down with me today.
(03:52) RS: Thank you, Alessandra.
(03:54) AB: I’ve been speaking with Ricardo Salgado, CEO of Loadsmart.
