October 10, 2017 | Written by: Siddharth Singh
Share this post:
As we look back at the banking industry’s past couple of years, they have been shaped by a push toward higher levels of digitization and banking automation. It is now clear that there’s more than one path to achieving this goal, depending on the departure points and appetite for risk and innovation. When I talk with clients specifically about automation, their stated goals range from capacity and throughput improvements, to short-term cost savings, to higher accuracy and innovation. Though most companies began their automation/digitization journey with finding a way to reduce costs and replicate human activities, automation as a path to artificial intelligence (AI) has become an important aspect of how to look at and transition to the future of work.
Nowhere is this more evident than in the banking industry, where many skeptics have become cheerleaders in less than a year. But I see an issue with both extremes. A lot has been written about skepticism toward robots and the new operational overheads, risks and controls that need to be enabled to allow such operations, so I’ll focus on the establishing a clear path to implementing AI technology.
When I talk with banks that are in the process of establishing an automation program, they often look to robotic process automation (RPA) and for it to solve all manual processing tasks. That said, there have been incremental evolutions that have included optical character recognition (OCR) technologies and some machine learning on the periphery, to deal with feature gaps in the toolsets. But in some ways, client expectations have evolved faster than any vendors’ ability to provide the right capabilities.
I concur with feedback I’ve received from many of my clients: automation programs now need to look at how to break out of the momentary excitement of robots automating basic tasks and create an automation ecosystem with a nerve center, smarter orchestration, process management and machine learning. This will allow various forms of advanced digitization, imaging, data extraction and machine learning, taking manual processes and bringing them online, and creating repeatable rule sets that are auditable and traceable.
Similarly, banking should start to see technology automation platforms begin to streamline and automate processes across major areas such as mortgage, trade finance, settlements, dispute management, risk and anti-money laundering (AML), as well as overall compliance—but also provide augmented intelligence capabilities for decision-based activities.
IBM predicts that in the next two-to-three years, standardized reference models will emerge around automation, where artificial intelligence (AI) and other foundational capabilities will coexist in the ecosystem with advanced machine-to-human and machine-to-machine communications capabilities that enable them to “talk” to one another—without building new apps.
On the payments side, for real-time payments and compliance checks, it’s becoming increasing convenient to use real-time payment systems (FTM) with Datacap, along with digital process automation and robot deployment to execute checks and reviews. Today, this may require a level of human review, but the combination of smarter digitization and extraction technologies—along with building knowledge basis using machine learning—is poised to provide review capabilities.
For capital markets, contract-based securities such as the repos and corporate actions, look at text-heavy content in securities contracts. Automation for broker/dealer or settlement/provider are well positioned to help extract business-context-driven information from these documents by using a combination of machine learning, natural language processing (NLP) and extraction technologies. These same technologies can also be applied to mortgage documents. Using automation in this area can help increase compliance, reduce complaints and provide better customer service.
Just as we have trail blazed major industry trends such as business intelligence (BI)/analytics, enterprise data platforms and more recently IBM Blockchain, IBM has already gotten the ball rolling with several banking entities by helping them establish a cognitive automation platform. While BI/analytics took over a decade to be fully adopted on an enterprise-wide basis, I believe this new cognitive platform will be adopted much more swiftly. At the core of this platform is what IBM calls a cognitive control room, which acts as the nerve center for the automation platform. This control room has access to business policies that drive the execution of various micro services, service orchestrators and scripted task bots, all through an integrated risk management and governance layer.
A great example which illustrates my thinking about this in an ecosystem can be best represented by IBM and HSBC coming together to automate trade finance. You can check out the short video here. This combined solution will use a host of technologies to help achieve end-to-end automation. I think there will be great reward in this, because it offers the potential to essentially turn a function into a utility—embed what has been thought leadership to become an essential part of day-to-day operations.
Connect with me on LinkedIn