Posted on FinExtra: https://www.finextra.com/blogposting/14667/making-ai-a-reality-in-banking 

 

Extensive automation of business processes in international banking using artificial intelligence, robotics and machine learning, which I’m going to collectively call robotic process automation (RPA), is set to transform the way banks and financial services firms do business – particularly in areas such as trade finance and similar complex banking products.

A wealth of data is generated day in, day out in the financing of international trade – from documentary credits to guarantees, loans and cross-border payments, and today there’s still a huge amount of manual processing involved in managing the process.

Clearly, as banks seek to reduce costs and increase efficiency, they’re looking to automate as much as possible.  Recent studies by Accenture have found that AI adoption could as much as double economic growth rates within the coming decades, with the potential to increase productivity by more than 50% compared to today’s old fashioned and legacy IT systems.

Robotic Process Automation (RPA) involves the creation of software robots that act as virtual workers who can be rapidly "trained" by business users in an intuitive manner. This is akin to how an operational user would train a human colleague. Crucially this frees up the limited and valuable skills of IT professionals to concentrate on more strategic tasks.

From anticipating supply-chain disruptions to predicting future customer behaviour and better managing stock, RPA-based platforms can process data immeasurably faster than their human counterparts, which is going to save banks a level of resources and manpower that was unimaginable as little as a decade ago.

RPA can also achieve significant operational savings in the area of compliance.  Ever-increasing compliance demands by regulatory bodies are taking up valuable resources in banking and trade finance which in turn, means passing the increased cost of doing business on to the customer.  In human terms at present, banks are required to have an army of compliance officers to hand, reviewing loans for international businesses. RPA can help solve this.

Of course, critics of AI and automation generally will point to the job losses that may be incurred.  Certainly, the nature of work will change.  Repetitive tasks will be automated, while more value-added tasks and exception handling will remain under the control of business and technology experts within the bank.  Once set up, the systems cannot be left to run unsupervised – the knowledge objects, components and systems must be monitored, controlled and maintained by human experts within the bank.  Similarly, there will always be exceptions and queries raised by the system that need to be escalated for manual intervention.

So how can banks start bringing Robotic Process Automation further into the mainstream?  For any banking system to be RPA ready, we must rebuild (or at least refurbish) from the ground up.  The software robots will need to have access to the hundreds of business processing rules at the bank - the relevant common objects and components, in addition to the specific product-related objects and components.  This will be required so that entire business processes, from the relevant portal system for the bank’s customers, through the mid- and back-office systems inside the bank, further on to the relevant business parties via the relevant channels (SWIFT, clearing houses etc.) around the globe, can be end-to-end automated to handle the relevant requests and responses to all parties, without human intervention.

To achieve this, banks need to scale back their reliance on legacy core banking systems and implement a modern IT platform built on international standards with all relevant business processing rules available.  This should include information and related business processing rules about the customers, the customer accounts, the customer agreements, the banks, the bank’s accounts, agreements with banks, agreement between involved banks, GL accounts, the products, the commissions, charges and fees, interest, the pricing rules, the clearing houses and clearing messages, the value dating, the relevant cut-off times, the business rules, and so on.

While this may sound challenging, the good news is that the technology exists for banks to get started today.  Some future-oriented vendors have already solved major elements of the RPA puzzle and are offering a route into this new world of handling banking products (in-house development would take 3-5 years and must be considered high-risk for banks).

Robotic Process Automation has the potential to make the world better.  We’re going to see more businesses able to access trade finance and other complex banking products – which is going to help businesses around the world grow.  Not only this, it’s going to cost less for businesses to gain financing too – all the while safe in the knowledge that robots are going to stop any maverick risk takers within the bank from stepping out of line, which is what regulation is meant for.  Sound tempting?  Get ready for the future, it’s already here.