Where can it be used?RPA can be used in a wide range of low end business processes that are done via application software, are high in volume, largely clerical in nature, are repetitive and rule-based, and do not involve a high degree of human discretion to be applied. Many examples come from the banking, financial services and insurance sectors, but there are also uses for this in operating ERP, and in healthcare.
Why is it of so much interest?RPA is of huge interest to corporate executives because of it’s potential to reduce BPO workforce costs by as much as 30%, while reducing errors (or improving output quality) and improving process productivity by potentially being able to work 24/7. The level of implementation difficulty is low, and implementation project lifecycle times could be in the order of a few weeks, which means that the risks of project implementation failure are low.
How is it implemented?RPA is implemented non-invasively, ie, it works with the human interfaces of existing systems, usually application GUIs or CUIs. If needed, RPA products can interface via web services or APIs, but by and large, it is implemented without the need for developer effort. The business process is analysed and mapped into the RPA software usually via a flowcharting-like modelling UI. It is then run on a trial basis, during which the model and business rules are tweaked until the desired output quality is obtained. After that the software is put into production under the governance of standard enterprise IT protocols.
Why is it of interest to the world of analytics?The very nature of RPA makes it a potential data source for both business as well as operational data in large volumes. As such it could be an enabler of further analytics, or it could be driven by analytics and machine learning. From both an operational or business perspective the data gathered while performing RPA tasks could be used by an interfacing machine learning system to fine-tune the operating quality levels of the RPA.
Where is the technology headed?
As a first step, RPA software exists as standalone products that with the enterprise computing’s machine human interface. However, with rapid parallel advances in big data technologies and the application of machine learning, it may be just a matter of time before machine learning and AI are used in RPA in order to improve the level of sophistication of tasks that it can be used for, while retaining its operating efficiency and cost advantages over a human workforce. Eventually, the BPM systems working behind the human interface and the RPA may also be integrated, and both may become embedded into the enterprise business applications. Perhaps in a few years from now enterprise business application software such as ERPs may largely operate themselves as a result.With the prospect of increased displacement of lower end white collar jobs, however, opinions are divided on what that would mean for employment rates, and whether the adoption of RPA would be hindered because of hesitations associated with this. While some industry leaders feel that the freeing up of large parts of the workforce makes them available to other labour shortage areas, others point out that in many environments the new jobs created may require completely different types of skills and skill levels. Either way, RPA is certain to be a subject of keen interest to many in the foreseeable future.