Technology

RPA is a Journey, Not a Destination


As companies begin their own RPA journey, the focus at first should be on defining their vision and then determine the right tool to help get them there.

©metamorworks/ISTOCK/THINKSTOCK

For some senior-level financial executives, the term Robotic Process Automation (RPA) produces visions of a utopian finance function. For others though, the mere thought of RPA brings on a major case of agita. 

According to a recent study by HfS Research and KPMG LLP, 55 percent of North American enterprises are looking at new opportunities available with RPA systems. Gartner predicts that by 2020, smart machines will be a top five investment priority for more than 30 percent of CIOs. So, regardless of which camp you find yourself in, the reality is that RPA is here now and will play a significant role in the finance function of the future. Companies will have bots as part of their digital workforce deployed throughout the organization. It’s important to note however, while RPA technology can speed up organizational processes, the bots cannot fix broken processes; at least not yet. Therefore, RPA is not the final destination but a technological step on a broader automation journey.

When Johnson & Johnson (J&J) hosted the FEI Committee on Finance & IT (CFIT) for their recent June meeting, Hugh Sheridan, VP, Business Technology Leader, Financial Services for J&J walked the group through several of the steps they have taken so far on their automation journey. According to Sheridan, “Process Automation is a strategic enabler for J&J and Global Services that is critical to achieving the long-term goals of improving the customer experience while reducing the costs of shared services.”

Their strategy is designed with a future state in mind where capabilities are optimized around near-real time actuals; intelligent automation, including RPA and Artificial Intelligence (AI); Data Science & Predictive Analytics; Driver-based Forecasting; Automate Planning; and Self-service Analytics. In order to achieve this, their strategy was broken down into four phases.

  1. Consolidate Work
  2. Standardize and Stabilize
  3. Digitize & Optimize
  4. Scale Digitization

As companies begin their own RPA journey, the focus at first should be on defining their vision – what do they want to achieve from using RPA – and then determine the right tool to help get them there. Automation Anywhere, Blue Prism and UiPath currently have a command on the market in terms of the RPA vendor landscape, but companies like WorkFusion, Pega, Redwood, NICE, and Kofax have been gaining traction and expanding their footprint. RPA strengthens and accelerates Shared Services’ ability to support customer needs. Based on a CEB analysis, there is a potential impact for a 15 percent capacity savings per employee.

Over the course of several RPA pilots, J&J found several learnings around five key areas.

  • Process – process validation and process stability are essential
    • How well do users follow the defined process and how much change is occurring in the process
  • Security – engage security team up front
    • Bots are digital people, they have identities, roles and access rights. How do you manage bots?
  • Environment – ensure due diligence is done to understand environment variation, infrastructure complexity, and source system performance that may impact bot performance
    • If your system performance is unpredictable the bots will be too
  • Playbook – utilize common enterprise strategies managed through CoPs to drive effective automation at scale
    • Bots are the 3rd choice; change the process, change the application, lastly bots
  • Agility – Don’t wait for perfection, develop MVP and iterate
    • Focus on good enough, 100% automation isn’t realistic

Automation potential exists across all general and administrative expenses (G&A) functions but is highest within Finance. According to McKinsey, up to 42 percent of Finance tasks could be fully automated. While an additional 19 percent are highly automatable. Transactional activities are the most automatable, but opportunities exist across most sub-functions. Some of the most commonly automated tasks include cash disbursements, general accounting operations, and revenue management. 

As the use of digital labor grows, an evolution of new capabilities will grow with this virtual workforce. It starts with basic automation and RPA which mimics human actions. This is a cheaper and faster step toward process efficiency involving screen scraping data collection, rules based business process management, and tactical toolsets to automate repetitive tasks. From basic automation companies will move to more complex automation and cognitive insights. This involves technology that mimics human judgment with basic learning capabilities for continuous improvement to quality and speed. Data input and output in any format, pattern recognition within unstructured data, and replication of judgment based tasks are all part of this phase. Finally, companies will move on to the use of AI for decision making. Here AI will augment human intelligence with self-learning rules continuously rewritten to improve performance using natural language recognition and processing, dealing with unstructured super data sets, and hypothesis based predictive analysis.

This is not merely a new technology; automation is a new workforce that dynamically integrates across business processes as well as our own human-capital workforces. It requires new ways of thinking – different processes, controls, governance, compliance, etc. Deploying automation requires a strong partnership across multiple disciplines. For J&J, they focused on establishing workforce automation as a service from the start. Deliver a holistic set of capabilities and a framework (business & technology) that will allow the functions to quickly deploy and maintain an automated workforce. They are starting with fundamental RPA capabilities, and then plan to evolve and grow along the ‘Automation Continuum’ as they mature and opportunities arise longer-term.