Mapping Your Automation Journey in Financial Services

Automation will fail to achieve most of its potential if it is treated as a kitchen gadget, a specialty, single-function tool applied to accelerate only a narrow part of a business process. In most industries, and in financial services especially, automation done properly is really a journey. Throughout this journey a steady stream of benefits are achieved as a broader and increasingly powerful multilayered stack of automated processes and analytics is built.

The idea is not just to have a bunch of separate enhancements, such as someone chopping faster in the kitchen, a more efficient process for taking orders online, and a more precise solution for organizing supply orders. Rather, the whole business should be seen as one integrated and automated process. While discrete parts of the system could be made better with traditional automation tools, newer tools such as robotic process automation, artificial intelligence and machine learning (AI/ML), and analytics can leverage basic automation to create a foundation that optimizes the system at higher levels.

What can this broader perspective of automation enable in practice? Consider these real examples of success:

• A large liquidity clearing business had been clearing its trades at the end of every day, which meant it only discovered defaults after the fact. This was a significant business problem. By identifying that challenge, they were then able to codify their clearing process to create a real-time clearing system. Not only does this flag defaults, but the system can also issue a line of credit to the customer, creating value for the customer and an entirely new revenue stream for the bank.
• Another bank identified compliance and governance as a critical area for enhancement to reduce their risk profile. By codifying their compliance processes, they’ve been able to create an automated system that picks up vital information from the flow of real-time trading to validate that transactions are fully compliant. It automatically tags, catalogs, and archives information in governance and compliance databases so that information is available on demand.
• AI/ML is excellent at fraud detection. By combining with robotic process automation, it is possible to steer customers to other applications across the environment to resolve issues and increase the protection of customers and institutions.

Naturally, these are the kinds of opportunities needed today, not just at the end of a multiyear deployment. The following four-pillar strategy makes it possible to reshape your business progressively, enhancing existing processes even as more extensive capabilities become available to you over time.

• Pillar 1. Fully Implement Basic forms of Automation. The foundation of intelligent automation is having the discipline to expand basic automation as widely as possible. Put simply, this is a mandate at the line-of-business level to reduce or eliminate manual processes wherever they can be performed as well, or better, by machine. The industry knows about such techniques; it just has to put them to work. Customer onboarding, portfolio management and many other labor-intensive activities are ripe for such business process automation. It should not take four days of people opening and closing multiple applications and filling out numerous forms in different systems, as well as manual back-end due diligence by people, to complete the opening of a bank account. These kinds of information-driven tasks are by default the kinds of things that machines are better and faster at completing.
• Pillar 2. Expand Basic Automation With Robotic Process Automation (RPA). RPA has two kinds of impact. It can fill gaps in basic automation that current techniques cannot address. It also can sit on top of current forms of basic automation and knit processes together in robust ways at a higher level. This type of automation addresses processes of a much broader scope that have lots of complexity and exceptions. By using RPA, you can push automation both deeper and wider.
• Pillar 3. Apply AI/ML To Process Information at Scale. Almost all forms of automation rely on expanding the data foundation. A byproduct of automation is expanded access to metadata describing the processes, but also more access to transactional data as well. AI and ML algorithm capabilities can make sense of this much broader landscape. AI and ML can surface new patterns that point the way to more automation, but also can detect anomalies that get in the way and require special handling. The result of applying AI and ML to automation is not only better automation, but a deeper understanding of the business.
• Pillar 4. Integration: the Last Mile of Automation. To have the greatest impact, all of the automation and insights gathered so far in the automation stack must be in essence productized and adapted to support high-volume processes. The last mile of automation means that you bring everything together and support the actions of a broad set of staff in a way that makes it easy to do a better job. Part of this is automation, but it also means creating advanced environments to recommend actions and to make it easy to process exceptions. This sort of automation can be applied across the entire business both for revenue-producing operations and for compliance and governance activities, such as daily FINRA analysis and reporting. In this stage, just as with AI/ML, the focus is both on automation and data.

Avoid Traps by Changing Perspective

Automation represents both opportunity and challenge for financial services. Some of the potential seems obvious; talk of making processes faster, more efficient and less costly is as old as business itself. That line of reasoning is also a trap. The tendency has been to think about automation in a contained and sharply focused way: as something applied in a discrete, narrow context. Automation should, in fact, must be viewed much more broadly, as part of the foundation of the industry.

Intelligent automation is that bigger notion. Instead of devoting resources to individual processes, the proposal is to apply automation at multiple levels in the business stack. It is predicated on understanding that the real objective is to make the entire business better, not to simply smooth out a single process.

These four interlocking pillars offer a novel way of understanding the automation stack. This perception makes it appropriate to bring automation business-wide to solve top-level business challenges in the financial services industry, rather than limiting its applicability to narrow tasks. Further, it makes it possible to recognize and capitalize on places where lines of business overlap. By integrating their applications with the underlying AI-powered data pipeline, the equities trading department can benefit from, for example, the information created by the customer onboarding department. Using this broad approach, it is no longer necessary to replicate processes multiple times to serve the needs of different parts of the business. Moreover, the use and feedback created by this system leads to the organic growth of new applications and opportunities.