The logistics industry is much like the rest of the world insofar as it is trying to grapple with AI’s role in both the present and the future. We know the technology is powerful and has the capacity to change the game in many ways.
But do we have a clear sense of its role, or of what needs to happen so AI fulfills its promise?
That was the conversation we had on a recent episode of Cargo Café with Chris Woods, vice president of brokerage at United Vision Logistics, about the future of AI and its impact on freight.
One critical issue is the fact that AI depends on the availability of quality data to excel in its role, and the logistics industry has some work to do there.
We're very fragmented in our data,” Woods said. “It lives in so many different places – in your TMS, inside your e-mail threads.
You’re looking for a PO or a proof-of-delivery or a conversation you had with a carrier or something like that. It’s in spreadsheets. It’s in ELD systems. We have to find avenues to bring it all together and let AI dissect what’s been going on in our ecosystem.”
A good example of the importance of quality data for AI is in ocean freight. And of course, it depends on how the AI is being used. Whereas AI is already well-positioned to help with tracking, the quality of data could be considerably more challenging when it comes to building out pricing structures.
“At what point is AI coming into the transaction?” Woods asked. “Is it at loading? Off-loading? Tracking? At the ports? If you use AI in tracking it can work wonders for you in ocean freight. But when it comes down to pricing, it’s critical to understand where you’re getting your data and how that can correlate to building out pricing structures and customer analysis.”
And while siloed data can sometimes be a problem, Woods believes it is very important to keep AI employees or agents siloed as they are assigned tasks.
“Once we build them to do so many things, how do you protect those decisions?” Woods said. “As you train them in carrier sourcing, or tracking, or updated notifications, it’s going to be important to maintain limitations on those bots while also protecting the data those bots have. I don’t want my AI bot telling one person that another person has a load that’s going to be delivered tomorrow. Why are you telling me that? It’s not my load. Keep the bot siloed in a function that keeps it simple.”
It comes down to this: AI is as good as the quality of the data that informs it, and as good as the decisions being made about what to have it do and under what circumstances.
Most industries are still figuring all this out, and logistics is no different. But one benefit of AI is that it’s helping logistics identify where it needs to not only improve its data but also its ideas for how AI fits into the bigger picture.
