In mid-August, Forrester principal analyst Sam Higgins and Automation Anywhere jointly hosted an executive roundtable with Chief Information Officers from the banking, financial services, and insurance (BFSI) sector. The discussion explored the shift in psychology, evolving roadmap, and changing prioritization surrounding intelligent automation (RPA, AI, and ML) investments as firms look for solutions they can deploy quickly to lower costs, support remote operations, and build resilience.
As a follow-up to the event, “Driving Business Resilience and Performance in Financial Services through Intelligent Automation,” Liam Kelly, director for financial services sector business at Automation Anywhere, spoke to Sam about the key insights from the event.
Higgins: Automation is a broad term and encompasses everything from physical robotics to the lofty desire for general-purpose AI. Prior to COVID-19, we saw a large degree of hype and, therefore, interest in artificial intelligence (AI), which in turn drove a focus on more-transformational or -aspirational automation projects. The pandemic has redrawn the automation landscape—at least for the next 18 to 24 months—as we see organizational investment priorities shift to automation solutions enterprises can deploy rapidly that lower costs, support the remote execution of business, and build resilience. Think solutions such as digital document process automation (DPA) for data ingestion and Robotic Process Automation (RPA) for processing. It is now a case of pragmatic solutions for today’s problems versus investments now perceived as potentially futuristic in the medium term.
Higgins: Decision-makers we survey report that their top benefits for adopting automation technologies are cost savings, higher levels of quality, and increased profit margins. When we look closer, it is technologies like RPA that contribute to human-machine coexistence that provide the most value, especially when automating routine tasks in business processes. Why? Two key reasons: Firstly, these technologies help minimize costs and risks associated with human error, freeing up employees to attend to more complex tasks. Secondly, these solutions automate routine tasks without having to re-engineer the underlying process. This means firms can implement these solutions quickly, and they provide a return on investment in a short time frame. Indeed, we often see firms using RPA solutions to unlock funding for longer-term automation investments.
Higgins: Firms that have successfully adopted complex intelligent automation—for example, RPA—integrated with machine learning or text-based analytics report that various business areas’ appreciation for these solutions mature at different rates. So while it is important to get C-level visibility and engagement, long-term success means growing an understanding of how and where automation can provide the best value with small demonstrations, proofs of concept, and production use cases. It also means expanding employees’ capabilities, such as developing up internal business analysts who are proficient in process assessment. Mature organizations have seasoned automation developers or service partners to design, code, and efficiently maintain automations. Finally, don’t forget to take areas such as legal, risk, and compliance on the journey. Build up their confidence by presenting to them the detail of how these solutions are deployed and operated.
Higgins: This has been a common question for the last couple of years. The answer is to start with simple processes in high transaction volumes—this is the RPA sweet spot, but we often see firms struggle with finding the right uses cases. The key to identifying which process to automate is to start with processes that aren’t too complex, so we recommend that firms use what Forrester calls the “rule of five.” The rule of five only has three rules, but they're all important. When initiating an automation program, apply tools like RPA to processes that contain tasks where there are fewer than five decisions made, five applications manipulated, and 500 keyboard actions per task. Once you identify these simple processes, prioritize those that represent activities that repeat often, are used by the largest number of people, and have the highest error rates. Errors are expensive to fix, but bots make fewer of them than humans do. As a result, we find organizations report stronger ROI through reducing manual task errors than just addressing manual tasks alone.
Higgins: This is a common concern for many firms that is supported by Forrester’s own data. In 2019, we found that over 50% of companies have fewer than 10 bots in production, with no shortage of issues. How do we find more good tasks to automate? How do we get the business building more robots but put the right controls in place? How do we build momentum? Automation can improve your business, but it affects internal systems, operations, culture, and management. So it’s not surprising that increasing momentum or achieving measurable velocity can be proven difficult. We recommend that firms start centrally, with a center of excellence or enablement (COE) or strike teams to drive RPA initially, but plan to federate responsibility to the business. Tune the operating model in these centralized teams, but then spread out design and maintenance functions to individual business units. Be sure to factor the creation of COEs or strike teams to govern RPA into the business case for long-term success.