Can artificial intelligence be trusted to manage logistics for an industry in which mistakes could be a matter of life or death?

Josh Medow, while an early adopter of AI, is aware of the risk. He is chief executive officer of Mercury Business Services, a provider of shipping services to the healthcare and life sciences sector. Coverage areas include biotechnology, pharmaceuticals, diagnostics, medical devices and equipment, and cold-chain shipments moving via national and local parcel couriers, as well as “white-glove” treatment for inside deliveries.

Mercury was actually founded in 1984 to handle the shipping needs of law firms. There’s little enough room for error when delivering critical contracts and other legal documents, but when e-mail took over that task, the company pivoted to healthcare.

Medow, who joined Mercury six years ago, is a longtime user of AI for business applications. Six months ago, he brought the technology in-house, with adoption of a chatbot built into the company’s transportation management system. Internal staff can query the AI assistant about route planning and other topics relating to transportation. Mercury regularly uploads an AI-generated document that guides its users throughout the process.

Included in the system is an AI status updater, which amasses messages from carriers to keep Mercury current on shipment arrivals, departures, pickups and other service details.

Medow considers AI to be highly effective for internal use, but he’s not ready to put it front of the customer. With so many chatbots suffering from hallucinations and coming to incorrect conclusions, he’s afraid that the technology will generate wrong answers. And the consequences of that happening in the medical field can be severe.

Problems that arise with medical shipments, such as a package getting stuck in Customs, are often too complex for automated systems to handle. When a customer calls with an issue, Medow says, the key to resolving it “is having an expert person [on hand].”

What would the AI have to show Medow, to prove its mettle in a customer-facing role? “That it can do better than my training,” he says. “It’s highly unlikely,” he adds, that a chatbot searching the internet can match or surpass the expertise of humans who deal with complex regulations and other nuances of medical shipping on a daily basis.

In the world of logistics, AI has many customer-service personnel fearing for their jobs. Medow, on the contrary, has doubled Mercury’s customer-facing headcount. As the company grows, he says, shipment issues become progressively more complicated. And customers are demanding to speak to humans at all hours of the day and night. To meet that need, Mercury maintains three shifts of logistics coordinators, based entirely in the U.S.

Medow continues to get pitches from AI providers seeking to augment, if not supplant, his human staff.
 “We’ve tried them all out,” he says. “They can do cool things — they can call a driver and ask where they are, but in real life I don’t know if that’s actually a needed thing. Our TMS gives a live GPS location of where the driver is.”

Medow acknowledges that AI is growing rapidly in sophistication and accuracy. The coming of autonomous vehicles and even cargo planes promises to have a huge impact on the sector, he says. But the need for humans isn’t going away anytime soon, especially in cases where a delivery requires special handling and crews to assist with uncrating and installation.

That said, “there are a lot of small things AI can do really well,” Medow notes, citing Mercury’s recent use of the technology to redesign its packaging and loading configurations in accordance with temperature controls and other requirements within specific lanes. “The geometric math is impossible for me to do,” he says. [AI] is really good at that.”