Standard Chartered, in partnership with IBM, has launched a service to automate the retrieval of information from trade documents.

Trade AI Engine is now live in 30 markets across Asia, Africa and the Middle East, with a further nine to follow shortly. The solution aims to tackle the perennial issue of cumbersome manual processing in traditional documentary trade through a combination of optical character recognition (OCR), machine learning and natural language processing (NLP).

The engine will convert non-digital shipping documents into a machine-readable format and classify them according to an initial pre-defined database. It will then ‘read’ the documents, capturing context from the data. The solution will learn as it goes, based on user re-classification and redefining of data elements, generating greater accuracy the more documents it handles. A spokesperson for the bank tells GTR that these unstructured documents will include invoices as well as transport and insurance forms.

“The Trade AI Engine enables multi-format document capture without templates and uses NLP to extract relevant information for review, as well as leveraging user interaction to improve text extraction accuracy and streamline the process with each cycle in real-time,” explains Likhit Wagle, vice-president of the financial services sector for IBM Asia Pacific.

While trade digitisation has made great strides in recent years, the industry is still bogged down by the weight of millions of paper-based, unstructured documents, often issued by various companies. Reviewing these requires a high degree of manual effort to ensure transaction compliance. By automating this process, the bank will be able to handle high volumes of diverse back-office tasks with greater efficiency and accuracy. Standard Chartered would not comment on whether this will mean a reduction in its back-office headcount.

“We process more than 36 million pages of trade documents annually, with over 200 million data elements for name capture and processing,” says Lisa Robins, global head of transaction banking at Standard Chartered. “The Trade AI engine significantly reduces the amount of time and effort in this review process while raising the bar on our controls environment, further enhancing our ability to play the connector role for our clients by facilitating transactions at pace with the growth of their businesses.”

Trade AI engine is currently available in India, Nepal, Botswana, Kenya, Uganda, South Africa, Zambia, Tanzania, Zimbabwe, Angola, Cameroon, Côte d’Ivoire, Gambia, Ghana, Nigeria, Sierra Leone, Mauritius, UAE, UK, Iraq, Bangladesh, Thailand, Philippines, Indonesia, Pakistan, Sri Lanka, Bahrain, Qatar, Jordan and Oman. The bank says it will shortly roll out the system in Japan, Brunei, Indonesia, Taiwan, Hong Kong, Singapore, Malaysia, China and the US.