AI Case Study
ING automates its compliance and sanction screening processes with optimal character recognition
ING has leveraged artificial intelligence to automate the otherwise traditionally paper based trade finance process. Using Conpend's ptimal character recognition (OCR) technology, it can automate the compliance and sanction screening processes. The system can automatically detect mistakes in trade finance documents and auto-correct them while also make them machine readable. With automating transactions, the company can achieve speed and visibility, with their ultimate goal being to increase transaction volumes and lower costs for their clients.
Industry
Financial Services
Banking
Project Overview
"ING’s guiding objectives when considering the applicability of existing technology to trade finance are to speed up turnaround times and make interaction with ING seamless for clients. It also seeks to streamline the bank’s processes to improve efficiency, reduce errors and – by substituting automation for repetitive tasks – free up time for functions that add value for clients. A further set of objectives is to make compliance with regulations, such as anti-money laundering, more straightforward and make it to easier demonstrate compliance.
A new solution introduced by ING, based on the Trafinas tool developed by fintech Conpend, achieves many of these objectives by using optimal character recognition (OCR) to scan trade finance documents and turn them into a machine readable format. The tool is capable of automating transactions, which improves speed and visibility. Two additional technologies – machine learning and neural language processing – will soon be added to the tool, improving functionality and performance further.
The solution will be able to automatically detect typing mistakes and auto-correct them in documents such as bills of lading and packing lists. Moreover, the use of machine learning means that when information on a document cannot be found in an expected location, manual intervention is only necessary on the first occasion: when similar problems arise in the future, the same fix can be applied automatically.
The use of OCR, machine learning and neural language processing will significantly lower the probability of errors in trade-related documentation. Furthermore, by digitising trade information, it will allow a bundle of documents that until now has been physically shipped around the world to be consolidated to a single electronic file. As a result, turnaround times will be cut, efficiency increased, and controls improved – a list of all parties to a transaction can be automatically compiled, which will facilitate instant comparison with sanctions or money laundering lists, for example."
Reported Results
"Ultimately, the new solution will enable ING to increase transaction volumes and lower costs for clients."
Technology
Function
Operations
Trading
Background
"Trade finance has existed for hundreds of years, if not millennia. As long as there has been trade between countries, there has been a need for companies and financial institutions to manage the risks associated with a process that, by definition, takes place in locations where the buyer or seller has limited visibility and in circumstances where they may not have a strong relationship with their counterparty.
Necessarily, the practicalities of managing those uncertainties are complex and often cumbersome. Even today the vast majority of documentation associated with trade, such as invoices, bills of lading or inspection certificates, remains paper based. As trade has become an ever-larger part of the global economy, and trade routes have lengthened, administrative burdens connected with trade finance have multiplied.
For companies involved in trade, the sheer volume of paper is challenging. Transporting, storing and accessing relevant information is difficult and increases turnaround times and costs. For banks, the use of paper-based documentation similarly results in complex processes and procedures.
Today, these burdens have been intensified by an increasingly stringent regulatory environment, with global, regional and national rules requiring an enormous amount of information about parties involved in trade. For example, banks are understandably prohibited from facilitating trade with countries facing sanctions. However, they also cannot be party to trading that involves a ship that at some point in its life has been connected with a country that has sanctions imposed against it. Paper-based documentation makes it difficult to achieve the visibility needed for effective sanctions screening."
Benefits
Data