The practical application of Artificial Intelligence (AI) by banks and financial service providers will be a hot topic in 2018. While seemingly a recent development, banks have in fact deployed AI to improve efficiency and lower costs for well over two decades. Starting with the use of Natural Language Processing (NLP) disciplines across several banking processes in the 1990s, the emergence of big-data and cloud computing drove the adoption of additional Machine Learning capabilities by banks. Robo advisors and fraud detection are two recent examples of AI applications in banking.
Despite the long history of AI deployment, we are now at an inflection point in the transformation of banking. This will see AI progress from its former role as a ‘nice to have’ enhancement provider, to a ‘must have’ essential facilitator of a mobile-first, cloud-first digital banking environment.
The deployment of AI in transaction banking has tended to be task-specific. Where AI will have a truly transformative impact on banking is its ability to solve highly complex and everyday problems in far less time than either humans or ‘traditional’ technology possibly could. This high velocity of ‘understanding’ is at the heart of changes underway —especially in two broad banking areas—financial crime compliance and process efficiency.
In an environment of three-day payment cycles and end-of-day bulk transactions, it is entirely possible for banks to throw yet more human resources at managing financial crime compliance obligations. This is costly and inefficient, but nevertheless the global headcount in compliance and alert management teams has increased to manage larger and larger workloads. Compliance and fraud detection today are very different from 10 years ago—the old monolithic national sanctions regimes are far more nuanced and complex. These are no longer ‘block’ embargoes, but are often now targeted at individual companies, individuals or commodity types.
With the global transition of banking and payments to a digital-first and real-time environment, compliance screening windows no longer span days or even hours. Today’s far more complex screening requirements must be made in seconds. It is simply not possible in this increasingly digital and 24/7 instant payment world to throw more people at the problem. The human mind lacks the cognitive capabilities to assess, investigate and evaluate the complex data body around a payment and reach a reliable judgment in that time.
A decade ago, AI delivered great benefits in financial crime compliance, lowering costs through fewer false positives and improved efficiencies. Today, as payments transition towards instant, AI is becoming fundamentally essential in providing the rich, context-aware capabilities to ensure real-time payments are made securely and with confidence. Traditional, static, rule-based technology is unable to do this.
The other significant area where AI is increasingly becoming an ‘essential’ business requirement in today’s real-time, digital-first banking world is in improving self-repair and routing across the transaction lifecycle—the ‘cognitive automation’ of payments through AI technology.
Despite significant investments in back-end processing and compliance, many banking systems—specifically the areas of payments processing, repair, routing and investigations — remain highly inefficient. The transformation taking place today is not the payments themselves, but the usage, customer expectation, integration and user interface demands.
Over the past two decades, the context learning and NLP capabilities of AI-based payments systems have been proven to dramatically increase straight-through processing rates. These also enable intelligent routing for cost efficiency, and result in the removal of inefficient manual interventions and repairs – though these AI benefits have mostly been the preserve of larger transaction banks. In a competitive environment, where instant payment validation becomes the norm, no bank can afford to ignore the power of AI to help realise the vision of full automation across the payments business in the digital age.
Self-repair, intelligent routing and false positive reductions draw upon a range of AI disciplines, including learning from human operator behaviour patterns. Using Machine Learning, AI systems are able to automate corrective actions and bypass human intervention for instant message repair and routing. This ‘cognitive automation’ will become increasingly important as banking systems adjust, adapt and self-learn to improve efficiency, meet customer expectations and reduce fraud.
As transaction banking enters a more open, real-time, digital-first, cloud-first environment, AI is set to transform the payments lifecycle process, becoming an essential enabler to provide the efficiency and security that banking customers will expect and require.
About the author:
Parth Desai is the founder and CEO of Pelican and serves on its Board of Directors.
With over twenty-five years of expertise in the practical application of Artificial Intelligence technology to payments and compliance, Parth has a thorough understanding of Payments, Securities, Anti-Money Laundering and Risk Management fields from both a business and technology standpoint.
Under Parth’s leadership, Pelican has grown to become a global market-leading company, delivering outstanding efficiency and product innovation benefits for customers in over 55 countries, and processing more than one billion transactions worth over US$5 trillion.
Prior to founding Pelican, Parth worked for Cognitive Systems as Research Scientist in New Haven under Roger Schank, a global pioneer of AI and Professor of Yale’s Artificial Intelligence Unit. Parth has a Master’s degree in Artificial Intelligence from Georgia Tech, a B. Tech. from IIT, Mumbai, and has participated in several senior management programs in Harvard Business School, Boston and INSEAD, France.