Artificial intelligence is having a dramatic impact on both shippers and carriers, helping to determine optimal pricing and promote transportation efficiency, says Tim Evans, vice president of sales with Loadsmart.
SCB: How is artificial intelligence transforming the transportation function today?
Evans: It’s happening on both the supply and demand side.
SCB: Let’s talk first about the supply side — how shippers get their pricing. How does an A.I. system come up with that? What kind of data is it drawing on, and how is it outputting that in a way that's more valuable than if a human were to do it?
Evans: What we’ve done up to now is human-based — you reach out to a broker, wait for them to get back to you with a price, then wait again for either that price to be accepted, or for the process to begin all over again. A.I. can input data from the market and provide an instant price, so a shipper knows within seconds what the price is going to be when they're ready to ship a load, and can move it instantaneously.
SCB: Does the system mix market intelligence with historical data on what you’ve been charged in that past?
Evans: Absolutely. It also accounts for seasonality in the marketplace. The system is predicting what’s going to happen in the marketplace, based on historical data as well as real-time input.
SCB: Is this of most use in the spot market? Or does the A.I. system also address contracted rates?
Evans: Traditionally it’s been used for spot rates, but we’ve also seen our business on the contract side grow quite a bit, as a result of being able to predict longer-term rates. Typically, shippers are overpaying to get that assurance from a contract perspective. What we can do through A.I. is leverage the technology to give you a more accurate projection of what that price is going to be in the marketplace over a longer period of time.
SCB: So shippers end up saving money on rates?
Evans: Yes. Brokers using this technology see the contracted rate become more aligned with what's happening in the spot market. A contract rate gives you the assurance that there’s a truck behind it. The spot market allows you to chase better rates. If you can close that gap by using A.I., you've got the best of both worlds.
SCB: Let's talk about the demand side. Where does the technology come in there?
Evans: Once you have a load and have priced it, you've got to make sure there's a truck behind it. The way to do that is use A.I. to project what the best carrier is for that particular load or particular moment. Who’s run that lane before? Who has the best performance? Who knows the facility of that particular shipper? All of this goes into a predictive model to make a recommendation of say, the three or four best carriers.
SCB: When you say “best” carriers, are you incorporating factors that are a little harder to quantify, such as the overall quality of the carrier?
Evans: Absolutely. You have different sizes of carriers, from the biggest to the mom-and-pops. All have different characteristics based on their size and ability to move a load. Sometimes the best carrier isn't necessarily the largest. Maybe a smaller one can cover a particular area of the country better.
SCB: How much acceptability has artificial intelligence seen in the marketplace, both from shippers and carriers? Do they trust it to come up with better solutions than their own human experts could?
Evans: It's slowly coming around. What will happen is that shippers and carriers with get enough experience to gain the confidence we have today in booking passenger flights on the internet. Today, maybe they’re using A.I. to handle 5-10 percent of overall loads. It's beyond the test stage; they're actually using it today.
SCB: Is it also a question of the system needing to gain experience, as with machine learning?
Evans: Absolutely. Every day, with each engagement, A.I. gets more intelligent because it's gaining information. Even when a price might be high, it learns from that to adjust the next time, so it’s gets closer to the market rate.
Enjoy curated articles directly to your inbox.