For all the headlines that artificial intelligence has fueled over the past year, the actual experience of many logistics and other business executives suggests that it remains more a subject of hope and hype than a practical tool in everyday use.

By one estimate, the application of AI to logistics is estimated to grow by between 12% and 60% from 2024 to 2026. The wide variance of that prediction, however, should give one pause. It suggests a deep uncertainty about the degree to which AI will actually take hold in the industry in the near term.

Even more questionable is the immediate value of generative AI, which creates “original” text, images, video and other data based on massive amounts of input scooped up from the internet and other media sources. A recent survey by BCG of more than 1,400 C-suite executives found 85% planning to increase spending on AI and Gen AI in 2024. Eighty-nine percent ranked it as a “top-three” tech priority for the year.

So much for aspiration. But the picture looks somewhat less rosy when depicting the current status of GenAI in business. Only 6% of companies said they had managed to train more than a quarter of their workforce in the technology to date. And 45% said they lacked guidance or restrictions on the daily use of AI and GenAI in the workplace.

Why the disconnect? Asparuh Koev suggests that the technology has been “sold wrongly.” The CEO of Transmetrics, which sells an AI platform for logistics planning, says the hype surrounding GenAI consists of “too much science fiction.”

Logistics executives buying into optimistic projections about the progress of AI assumed that robots were going to take over human tasks, Koev says. “In reality, at least in logistics, this is not possible.” There’s no question that the logistics sector can benefit from AI, but when it comes to the actual delivery of services, the industry remains decidedly human-dependent.

The logistics challenge consists, at least on its face, of “a very simple problem,” Koev says. “It’s being able to deliver the goods on time, without damage, and doing it cheaper than your competitors. Nothing else matters.” 

AI comes into the picture as means of determining how logistics service providers can derive extra revenue from the most profitable customers, based on the resources available within a particular market geography. 

The problem is how to achieve that objective while making use of the massive amount of data that’s available to logistics providers today. In a time when every shipment is repeatedly scanned, and warehouses and carriers are rapidly embracing digital processes, the challenge lies in plugging AI into that information gusher. “They’re failing with that,” Koev says. “They’re trying to run AI with minimal data sources.”

One reason, he says, is that all of that available data is trapped within multiple information systems that aren’t properly integrated. “You may have a million data points,” Koev says, “but if you don’t connect them properly, there’s nothing you can do with that huge amount of data.”

Even when the data is put to constructive use, AI has limits. The technology can generate recommendations about how best to serve customers, but human intuition remains the ultimate determining factor — at least for now. Koev says AI is still a long way from being trusted for that purpose. “When you have an AI application, the user isn’t satisfied that it’s better than human. It has to be 100 times better.”

Further hampering the development of AI is the slow process of automating — or, in modern-day tech-speak, digitizing — manual processes for managing logistics. That’s a critical precursor to the implementation of AI, which “cannot operate outside of the digital world,” Koev says. “If trucks aren’t moving, or it’s not clear where they are, and data is not available — what do you expect an AI to do?”

Ultimately, he believes, AI will become sufficiently mature to take over many key business processes, but humans will remain in the mix. He likens the future world of logistics to present-day automated pilot systems deployed in aircraft. For the most part, they can operate without direct human intervention. “But we’ll never allow an airplane to fly without two pilots in the cockpit.”

Similarly in logistics, “you are always improvising,” as customers change their minds about how they want product delivered, or unanticipated disruptions force last-minute changes in routing.

Koev’s one exception to his insistence on the permanent presence of humans in logistics is the prospect of fully automated, self-driving vehicles. AI will play a crucial role in ensuring their safe and efficient operation, he says, “but I don’t believe this is going to happen anytime soon.”

What will happen is that AI, and GenAI specifically, will eventually come to play an indispensable role in logistics and other industries. BCG believes GenAI will make it possible for organizations to “increase productivity, enhance efficiency and effectiveness, boost revenue, and build a long-term competitive advantage.”

It's just a question of when.