Robotic process automation (RPA) refers to software that can be programmed to perform basic tasks across a range of platforms and applications. Designed primarily for office-based functions, these tools have the ability to perform several tasks in order.
Listed below are the top robotic processing predictions, as identified by GlobalData.
RPA will continue to be a ‘way out’ for large incumbent banks that need to limit the cost and process complexity of legacy systems but don’t have an easy way of exposing APIs or the stomach to migrate or decommission core legacy systems. RPA will work as a low-cost layer that can sit on top of all existing banking applications.
Instead of being forced to rely on complex codes and IT department intervention, front-line banking employees will be able to automate their own work, once trained properly. RPA will begin to disrupt business process outsourcing models, as it will provide a lower-cost, higher-productivity model.
To optimise spend, financial service providers will expand their goals beyond automation for isolated tasks. Combining RPA with AI components such as text analytics, conversational intelligence, or decision management based on machine learning will allow banks to attack broader use cases.
To support these goals, process discovery analytics, the ability to scale beyond a small number of production robots, central orchestration, and platform openness will grow in importance. In turn, leading vendors, such as Automation Anywhere, will make strategic acquisitions, investments in AI research, and partnerships as they jockey for position.
2020 will see widespread use of RPA in mortgages, where applications can involve over 700 pages of agreements and legal documents, which are not only expensive and time-consuming to process but also error-prone. Roostify in the US uses machine-learning techniques to improve accuracy by recognising patterns in textual analysis following the use of optical character recognition to ‘grab’ data from paper documents.
Singaporean bank OCBC has reportedly been able to use RPA to reduce the amount of time required to re-price home loans from 45 minutes to just one minute. The process checks the customer’s eligibility to have their home loan re-priced, recommends re-pricing options, and even drafts the recommendation email.
Ultimately, 2020 will bring home the notion that RPA will not prove a process ‘silver bullet.’ It cannot fix processes that are broken in the first place. Simply layering it on top of suboptimal banking processes will neither address nor fix the root cause of organisational process problems. Most banks struggle to merge different legacy systems into a single workflow for RPA.
RPA for basic customer service will continue to reduce costs with minimal broken glass in terms of customer satisfaction. When combined with AI, RPA can even deliver more personalised customer service. But new digital banks will continue to overstep the mark, with many Revolut customers, for example, complaining about the over-reliance on RPA and how anonymous, impersonal, or inappropriate interactions can be.
This will see increased focus on channel transitions, handovers, and escalation procedures so that customers exit bots as soon as their language or service issues prove too involved or emotional.
This is an edited extract from the Banking & Payments Predictions 2020 – Thematic Research report produced by GlobalData Thematic Research.