AI Case Study
HSBC tests automated information extraction system for global trade contracts
HSBC is piloting an automated ledger technology which extracts key information about buyers, sellers and other logistics from trade finance documents. The documents are first digitised and then IBM's technology extracts information which is automatically integrated into the bank's processing systems. Automating a currently labour-intensive data capture process to increase accuracy and efficiency by reducing need for human review.
The pilot is being used for global trade and shipment contracts. "The IBM analytics technology, initially in English language import and export bills — about a third of the total paper volume — will enable HSBC to jettison much of the mundane clerical work and increase accuracy. The IBM technology that can extract information from non-standardised papers and feed it straight into bank processing systems."
Pilot; results not yet available
"A single contract could run to 100 pages. Everything has to be in triplicate and a shipment of, say, 500 shirts in various colours and sizes — each taking its own page — quickly stacks up. For now, it falls to clerks to pull out relevant information: from dates and addresses to the name of the ship’s captain. The typical transaction requires 65 pieces of information to be pulled out of an average 40 pages."