Authors’ note: This is the fourth of a five-part series, in which we discuss the ways banks are adopting robotic process automation (RPA). In this installment, we explore what banks should be thinking about when moving from proof of concept-based, siloed, task-level automation to overall operations transformation and how platform vendors are working to support these efforts.
As we explored in Part 3, many banks are implementing point solutions to automate key parts of processes without significantly overhauling their underlying systems. The approach is to look for quick and low-cost opportunities to achieve short-term wins while major initiatives are underway. RPA helps bridge that gap by providing relatively inexpensive and even temporary alternatives. However, the next wave of automation will be more strategic and permanent in nature.
Most banks we work with today are still maturing in their use of task-oriented automation with a single automation tool. Automation solutions are typically applied to routine, repetitive, human-assisted tasks, and they often deliver 20% to 30% cost savings, based on our experience working with banking organizations. A few banks, however, are further along the maturity curve. Because their task-level solutions have performed well, they’re moving toward enterprisewide automation using cognitive technologies.
Certainly, we’re all excited about the potential that cognitive automation presents, but before jumping too quickly into these emerging technologies, it makes sense to explore the complexities associated with these technologies, set realistic expectations about ROI and understand how automation tool vendors are charting their specific product roadmaps.
Consider the following:
- End-to-end process automation incorporating cognitive. Some banks are working to orchestrate multiple task-level automations across multiple processes. This approach often combines repetitive, routine tasks with other tasks in which conclusions need to be drawn while the process is being executed. These more complex tasks are typically not human-assisted, and they usually require cognitive processing to generate insights. The potential cost savings and business benefits of these end-to-end automation projects can be higher than simple task-level automation due to the compounding effects of multiple task-level automations, the unassisted nature of the automation, and the insights that cognitive technologies can provide.
For example, document intake for know-your-customer (KYC) processes often become a bottleneck when banks try to automate the entire process. By applying machine learning (ML) to intake processes such as those driven by optical character recognition (OCR), banks can improve the accuracy of meaningful data extracted from scanned documents to ensure they are fed correctly through the remaining automated KYC processes.
- Technologies required. Simple task-level automations, using a single tool such as those from UiPath, Blue Prism or Automation Anywhere, are the most readily available options for conducting quick proofs of concept (PoCs), validating organizational readiness for automation and demonstrating business value.
Conversely, more complex end-to-end process automation applications usually require emerging technologies, such as machine vision, natural language processing (NLP), ML, adaptive alteration technologies and chatbots, among many others. No single product vendor currently provides all of these in one platform.
- ROI comparison. Task-level automations produce fairly quick ROI due to the low initial investment and licensing costs associated with a single tool, as well as a short timeline to implement, usually a few weeks. The cost of staff with task automation skills is significantly lower than what’s associated with cognitive automation. Any developer or operations professional with a strong understanding of applications programming, such as Java, .Net or VBA, can be trained on task-level automation tools quickly. In addition, one license typically provides 24×7 availability of the bot, and the incremental cost of automating additional tasks is lower, so the ROI increases along with the number of tasks automated.
ROI for complex, end-to-end process automation may not be as quick. Since there is no single vendor for all cognitive solutions that may be used, banks will need to invest carefully in multiple tools. End-to-end process automation can also take months to implement because of the time required to justify the business case, identify and onboard new tools, and hire skilled staff. Cognitive automation tools often require highly skilled resources and data scientists. Also, cognitive solutions need massive transaction volumes and enough time for the machine to learn so it can provide meaningful insights, which adds to the project’s timeline. Licensing costs for cognitive automation tools are several times that of task-level automation tools.
All of this combines to make the initial investment higher and break-even period longer. That said, if done well, the investment can offer a significant and enduring advantage in terms of cost or customer experience.
- Vendors jumping on the cognitive bandwagon. With the hype surrounding automation, many startup companies and established vendors are offering automation solutions. Approximately 30 major players already occupy this crowded space – more than double the number three years ago. Established vendors, such as Pegasystems, NICE Systems and Kofax, have incorporated automation into their existing platforms.
Total investment worldwide in automation platforms has increased dramatically. Even though tools are still maturing, and most remain focused on task-level automation, some – such as WorkFusion and UiPath – have defined ambitious roadmaps for their cognitive solutions.
With this fierce competition, tool vendors are trying to differentiate themselves through unique solution features. Banks need to be careful with this because some features, while interesting, may not be a necessary part of the banks’ own automation roadmap, and choosing that solution may create vendor lock-in. We suggest that businesses carefully evaluate the feature roadmaps of various product vendors to assess alignment with their own automation journey and the degree to which they are truly open from an architectural standpoint to enable integration with other tools.
Depending on their individual automation roadmaps, banks may need to put together different cognitive components themselves since no single platform offers all cognitive capabilities. For example, some of the chatbots and machine learning tools that banks will need for end-to-end automation are available through Amazon, Google and Microsoft platforms. Other tools, such as NLP from Nuance, are not necessarily automation products themselves, but they can play a key role in end-to-end automation by integrating with automation platforms.
Still other vendors choose not to incorporate ML into their automation tools, instead providing integration points for open source cognitive computing solutions, while others, such as WorkFusion, embed cognitive capabilities in their platform to make it possible for operations teams to automate repetitive tasks without help from data scientists. A bank will need to invest in developing the required expertise to evaluate and integrate these options or rely on an integration partner to help achieve the desired outcome.
The Automation Future Is Bright
As banks move along the automation maturity curve, many opportunities for process improvement, cost savings and data-driven insights await them. As they master task-level automation and gauge RPA’s short-term potential, it will be important to keep an eye on the horizon.
In the final installment of this series, we’ll recap the key insights from our first four blogs, which focused on the “seven deadly sins” that can derail RPA adoption (Part 1); “seven virtuous steps” to avoid these deployment pitfalls (Part 2); and where and how banks are adopting automation to maximize ROI (Part 3). Part 5 will also offer additional ideas to help banks take a more holistic and impactful approach to their automation journey going forward.
Sriniketh Chakravarthi, who leads Cognizant’s North American Banking and Financial Services business unit, contributed to this blog.