Robotic Process Automation – AI under the microscope

RPA has emerged as a must-have for many financial institutions, but what is it, and how can it be applied?

Artificial Intelligence (AI) has been a hot topic for some years now, rising up the agenda in many board rooms due to its increasing media coverage and clear untapped potential. Under the umbrella of AI falls a number of methodologies, one of which is Robotic Process Automation (RPA). RPA is emerging as a must-have for financial institutions, with its software revenue growing by 63% in 2018. However, it is important to understand what differentiates RPA from other AI methods and where it can be useful to implement the technology.

The enthusiasm for automation transcends the financial services space, but the deployment of RPA, if done effectively, holds a number of advantages for financial institutions. These include reducing the time spent on administrative and transactional activities and allowing workforces to dedicate more time to other, more valuable tasks.

RPA is mostly process-driven, acting upon the information at its disposal to make deductive decisions. It is valued for increasing productivity and ensuring that tasks are completed accurately. Often characterised as the brains and the brawns, RPA will carry out rules-based activities based on set criteria but will not be able to ascertain whether the results of its actions are good or bad. As a result, RPA deployment, and by extension wider automation, is not guaranteed to succeed without prior knowledge of where problems and inefficiencies lie. Nevertheless, the data that RPA collects can provide the basis for a more intelligent approach via machine learning and AI.

A first step towards intelligent automation is RPA that can store data upon which machine learning can be applied. In order to conceptualise its role, dividing the automation process between “obeying” (RPA), “learning” (Machine Learning), and “thinking” (AI) allows us to clearly situate RPA. RPA, while “obeying” its prescribed rules, can begin to store information as instructed, which can become a database for “learning”. Machine Learning can then be applied to recognise patterns in the data, enabling AI to step in and make intelligent decisions which feedback to alter the RPA’s rules. In this process, the value of RPA is to act as the first step towards intelligent automation.

Efficiency gains and cost savings are the big wins for RPA systems. Near 24/7 availability and higher accuracy make any application of robotics an appealing proposition for businesses wishing to reduce their costs. Estimates suggest that RPA would cost about one-fifth of a worker and one-third of an offshore employee, while the speed at which RPA can be deployed leads to estimates of less than a year to recoup initial investments.

With the financial services space being full of repetitive tasks on a massive scale, it is not hard to understand why RPA is seen as an appealing prospect. According to reports, 89% of accounting operations and 72% of financial controlling and external reporting operations are highly automatable. The suggested tasks that RPA can undertake in the financial services space often pertain to compliance, onboarding, and verification, though there are many others across various business functions, including:

  • Bank Reconciliation Processes
  • Sales Ordering and Invoicing
  • Financial and External Reporting
  • Payables Management
  • Receivables Management
  • Financial Statement Closing

Source: EY

As compliance becomes an increasingly pertinent everyday task for financial institutions, leveraging RPA capabilities effectively could alleviate some of the associated burdens and concerns. Effective RPA technology could also lead to frictionless customer experience in the realms of authentication, customer queries, and many other B2C interactions, with B2B scenarios surely set to follow.