“Generative AI is the defining change of our careers — maybe our lifetimes. It’s like the internet revolution but compressed into a couple of years.”
That was the contention of Jonah McIntire, chief product and technology officer of Trimble.
At a recent Trimble roundtable, leaders discussed how artificial intelligence is reshaping logistics — from back-office email to freight optimization — and what it takes to use it well.
AI, especially the new generation of generative AI, can be used in everything from pricing emails to route optimization. But is it all just hype? Or is it already quietly reshaping how freight moves?
The increasing connectivity between vehicles, software, and more is the base on which artificial intelligence is being built.
“The next value unlock in those connected systems is leveraging the rich amount of data that is passed into and out of these systems to create some outsized outcomes,” said Michael Kornhauser, Trimble sector vice president, transportation and logistics at Trimble. “And that’s where AI can really come in.”
How Common is AI?
Artificial intelligence in general is not new. Machine learning and AI algorithms have worked behind the scenes for years, even decades. But the generative AI, large language models such as ChatGPT or Claude, are “a different best,” as McIntire said.
McKinsey & Company, he noted, has been doing surveys about the adoption of AI among businesses going back nearly a decade.
“If you think about what AI was in 2015, it was things like predictions and forecasts or categorizations, those sorts of things, and not generative AI,” McIntire said.
McKinsey’s research found a rapid increase in the number of organizations using artificial intelligence in at least one business function between about 2017 and 2019. But the number remained fairly flat, around 50%, until the introduction of generative AI a couple of years ago.
Since then, the number of companies using generative AI have increased rapidly – and at the same time so has the number of companies using traditional AI tools.
Looking at Trimble’s customer base, McIntire said he’s been at customer roadshows and asked the audience how many people are using generative AI in their business, and nearly every hand went up.
“They may be using it for use cases that seem pretty quaint, like crafting a better email to a customer to inform them of a price increase,” he said. Nevertheless, “A lot of people are using a technology that wasn’t there two years ago, so the technology adoption curve is very high.”
Real-World Use Cases: Where AI Delivers Now
The Trimble experts outlined a number of use cases where generative AI can deliver for trucking and logistics companies, both through better efficiency and bringing in more revenue. For instance:
Deal Verification: Artificial intelligence verifies companies and their representatives in Trimble’s freight network, replacing tedious human checks and preventing fraud. The AI agents are not only faster, they’re also more accurate, McIntire said.
“The work is kind of grinding, right? This is the kind of thing that AI agents are good at. It’s diligent, tireless inspection of facts against sources of truth.”
Email-to-Action: AI can convert incoming freight quote requests into structured data and responses. Say a carrier has a customer who sends them an email asking for a price quote on a shipment that is not its usual freight.
“Right now, there’s human glue in most companies that would sit there and structure that over into something else, and then maybe look up cost structures and go try to find what the cost would be, and then decide how much margin to add on top of that, and then reply back,” McIntire said.
“Those are things that are fully automatable for some of our customers now, because we use generative AI for that.”
Negotiation Tone: AI helps sales reps choose the most effective communication style for pricing responses.
Do you mirror their tone? Counter in a very direct, even demanding tone? Or do you counter in a very, ‘I’m so sorry I’m unable to accommodate you’ tone? AI can run the numbers statistically to tell you which is most likely the better tone in any given situation.
Shipment Visibility: AI analyzes real-time data to flag potential disruptions and suggest actions.
What happens when a shipment begins to deviate from the plan? At what point do you get concerned and contact the driver? Maybe the road is closed because of an accident or the weather is slowing them down.
“A lot of data goes into what seems like a very straightforward decisions, but it is an example of artificial intelligence,” McIntire said.
He predicts that in the next year, AI will take this a step further by contacting the driver if it determines it’s needed and talking to them in natural language and voice.
Research-on-Demand: Generative AI tools prepare briefing documents for sales or recruiting in minutes.
Say your sales department is looking for new shipper customers. You can ask generative AI tools such as Gemini, ChatGPT or Claude to write up a research report on companies you want to contact, telling the AI agent what information you want included.
“You’re essentially preparing a briefing for the salesperson or the business development rep so that they can sound knowledgeable and prepared when they make contact with that customer,” McIntire said. “This is something that would have taken them two or three hours to research and create, and now it’s available to them in maybe five minutes.”
Philipp Pfister, sector vice president for Transporeon, said he sees AI in essentially two buckets: Efficiency gains and generating more business. Using it to help in deal-making can do both.
“Generative AI can help you to be faster in getting deals and help you to lower the amount of people that need to deal with the exact same thing, so you have basically both the efficiency gain as well as more business for your own trucking business.”
Battling Bad Data and AI Hallucinations
AI is only as good as the data it’s fed — but that’s improving.
When Trimble works on integrations, Pfister noted, it looks ahead and works with customers to make sure they have clean data that will be “AI-ready” when the time comes. There’s even a dashboard to show customers how good their data is.
“If you have a way to validate data as being good or not,” McIntire said, “You can go from very unclean data to very good training data because you essentially filter out the bad data.”
McIntire added that generative models today are smarter. With the right programming/training, they often can spot and flag bad data that older systems might have accepted blindly.
“There was no common sense at all in that generation. It was doing advanced calculation, undoubtedly, but it was doing advanced calculation without really any sense of what good or even real looks like.”
That likely explains why AI adoption was pretty flat between around 2017 and 2020, he suggested. “People found use cases, but they also found the limitations of those use cases.”
AI “hallucinations,” such as we’ve seen in recent media coverage of non-existent books put on a summer reading list or non-existent research papers in a government report, can be avoided with the proper programming up front, McIntire said.
“You establish ground truth validation techniques upfront,” he said, creating a test and repeatedly testing your code against it until it’s right.
He compared it to having a spell-check or a human editor reviewing a document for errors.
Advice: Where to Start in Your AI Journey
Because this is all changing so rapidly, many businesses are taking more of a wait-and-see attitude toward adopting AI.
“I think that the fact that it’s very unknown and very new is the biggest challenge,” McIntire said. “There are not known best practices that one can simply adopt. It’s a domain that largely didn’t exist until November 2022, and it since then has changed with a pace that people are not yet used to.
“Most of the leaders we talked to, our customers, they want to do something. They don’t yet have a stable best practice that they can just adopt.”
Pfister noted that some aren’t even digital yet — “they’re still using pen and paper” — and need to modernize first.
“We need to acknowledge everyone has a somewhat different starting line and the finishing line keeps on moving right now,” he said.
For fleets that want to stick their toe in the water in exploring today’s generative AI technology, McIntire suggested a subscription to an AI program such as Claude, Gemini, or ChatGPT for key staff.
“I would start with back office staff who are doing research or doing some sort of communication-focused work,” he said.
Don’t Adopt AI Just to Adopt AI
The Trimble panel emphasized that when adopting artificial intelligence, it makes sense to first look at where a company’s operations can be improved and how AI can help do that.
“We’re in this age right now where every website has an AI assistant sitting over on the side, awkwardly, waiting to be summoned for tasks that they do with varying degrees of competence,” McIntire said.
“And that is not our strategy. Our strategy is to pick a problem the customers have, and then solve it.”
He compared it to trash collection. Trash collection is actually fairly complicated logistics. But for the customers getting their trash collected, it’s not something they have to think about. It’s just there.
“That’s the kind of intelligence we’re trying to bring into our applications, where it’s just ubiquitous. It doesn’t constantly wave at you and try to show you how cool and snazzy it is.”
Credit: Source link
