Cody Williams, director of solutions consulting with e2open, clues us in on the progress and long-term potential of artificial intelligence in global trade management.
AI “absolutely is” making its mark in global trade management today, says Williams. He notes the existence of multiple use cases, in which AI training models were successfully applied to such key processes as document processing for creation of structured data sets to facilitate entry processing, and classification of imported goods under the Harmonized Tariff System.
It’s a relatively recent development — the use of AI and machine learning has spiked over the last 12 months, Williams says. Over that time, GTM teams have had to address concerns about compliance with a new technology, to make sure that the AI isn’t “hallucinating” or applying incorrect values to imported goods. Even in that brief time, however, the resulting impact on classification and document intelligence has brought “huge benefits for importers.”
The recent tariff turmoil cries out for an efficient means of keeping pace with changes in duty levels and trade rules. The application of AI to HTS classification has made it possible to “engineer” tariffs, so as to bring in product at the lowest legally allowable duty. In the process, the technology is actually dictating how a product is created, and how its various components come together.
There’s still a need for humans to validate the conclusions of AI, Williams says, with particular emphasis on ensuring that the right data is being used by the training model.
The machine learning component doesn’t take that long to “learn” the fundamentals of a company’s trading system. The biggest challenge, Williams says, is obtaining the data from multiple systems — product lifecycle management, enterprise resource planning, customs entry — and ensuring that it’s accurate.
In the future, he says, AI will enable an unprecedented level of collaboration among all of the functions that make up a complex trading organization.