Banks today are driven by the dual objectives of driving business innovation to deal with the threats created by the fintech players and the ‘digital only’ neo-banks, and creating cost optimisation through operational efficiency initiatives. A key lever in driving operational efficiencies and cost reduction has been the deployment of Robotics Process Automation (RPA) solutions. A RPA solution refers to the deployment of “software robots” that can be configured to create a virtual workforce that can mimic the human actions needed in many of the banking processes. The robots need a one-time “training” through a process flow mapping exercise but, once trained, can carry out the repetitive tasks in a process at much higher levels of speed and accuracy than a human can. RPA solution providers such as Automation Anywhere, BluePrism, Ui Path and Pega RPA (earlier OpenSpan) have seen large scale adoption in the banking and financial services industry. Most of the large banks have hundreds of bots automating a wide range of processes from the front office to the back office.
Making RPA more intelligent
In the first wave of the virtual workforce adoption in banks, RPA providers and banks have focused on using robots for high volume, repetitive tasks (referred to as deterministic RPA) and freeing up the human bandwidth in the operations departments to focus on higher value tasks. But as the benefits realisation from the deterministic RPA implementations begins to taper out, banks must start looking at cognitive RPA solutions and the deployment of Machine Learning (ML) and Artificial Intelligence (AI) in their process automation initiatives. This new wave of process automation would be focused on increasing the level of intelligence and making the robots more autonomous. This intelligent virtual workforce would be a lot less dependent on human interference and be able to perform more complex tasks that need decision making beyond the “if-else-then” kind. Some of the solutions that offer cognitive RPA and AI/ML solutions include Workfusion, IBM Watson and Fast Forward Labs. For banks to realise the full benefits of cognitive RPA, however, they must combine it with technologies such as speech recognition, natural language processing (NLP), ML and OCR (Optical Character Recognition) / HCR (Handwriting Character Recognition). A typical use case can be trade surveillance where call monitoring using voice to text conversion first and then using cognitive pattern recognition can be used to detect trade fraud. Other process areas where cognitive RPA and the related technologies can add a lot of value are the ones that involve many contracts and invoices: e.g., trade finance and derivatives trading.
Managing the ROI conversations
As we move into the realm of cognitive RPA and ML/AI solutions, ROI realisation would take longer. The business case therefore needs to move beyond just the immediate cost take out. The following are parameters to consider in coming up with the business case:
Increased accuracy with lower cost: Robots with cognitive ability can play the checker role in addition to the maker role (realised using deterministic RPA) in many of the maker-checker functions in banking operations. This can lead to redeployment of more senior bank staff such as supervisors and therefore higher cost savings eventually
Better analytics: Using a combination of human plus machine capabilities for judgement based processes can drive more informed business decisions.
Reduced risks: Cognitive robots deployed in the regulatory space, for example, can reduce risks related to reporting errors by minimising the human factor
Cognitive RPA and the use of ML/AI in banking process automation is set to see an exponential rise in the next couple of years. Banks that lead the adoption will see much better ROI benefits. The time to make the virtual digital workforce more intelligent is now!