AI assistance for automated order entry

How can logistics benefit from Artificial Intelligence (AI)? Based on a proof of concept at a. hartrodt Germany, the IT service provider msg systems presented this at the International Supply Chain Conference last week in Berlin. In the pilot project, it was possible to prove that AI and Machine Learning can ease the high manual effort for order entry. "We saved up to seven minutes per standard order," reports Stefan Goestl, Airfreight Manager at a. hartrodt in Munich. He provided anonymized data from legacy systems for his department with 14 employees.

Orders by e-mail with PDF attachment

According to Goestl, 90 percent of orders in Germany are received by e-mail with PDF attachments: "Depending on the size of the shipment, this can result in a lot of time spent on documentation." Freight forwarding employees classify the mails, create the order, extract specialized data, reconcile master data and often have to make follow-up phone calls. "AI enables a significant increase in quality, and this gives us more time for individual customer care," says Goestl.

Target clearly exceeded

In the AI prototype, 100 percent of the orders were recognized and pages classified without errors, and 97 percent of the data was extracted correctly. Goestl was "positively surprised" by this, and the target of 90 percent was clearly exceeded. With automated order entry, freight forwarding employees only need to check incomplete or unclear orders – the AI assistance learns and improves through user feedback. Data is automatically transferred to the CargoWise One supply chain platform. "There has never been anything like this in logistics on this scale," emphasizes Philipp Spring, Manager Business Support, IT at a. hartrodt in Hamburg.

According to Spring, a. hartrodt is pursuing a "multi-year project" with AI assistance, which can also be applied in other divisions. "We have to move with the times and align ourselves with the future. For this, IT is the key pillar," he points out.