Robotic process automation revisited: What's really going on?
- 08 February, 2019 07:30
Robotic process automation (RPA) is a recent technology based on software robots that are configured to carry out business processes previously done by people. The software acts as a virtual worker, programmed to carry out operational procedures (Wilcox, Leslie, and Mary Lacity. 2016. Service Automation Robots And The Future Of Work)
Software robots, or as now known as bots, interact with structured data, capturing, interpreting, and manipulating it across digital systems to trigger responses and outcomes to support business processes. RPA service providers have been promoting the benefits of this technology since its initial implementation, claiming that RPA reduces costs while increasing service productivity, accuracy and speed.
This study was carried out to examine whether RPA has indeed delivered these benefits to clients as well as shed light on the state of this emerging technology and its service providers.
We identified 155 articles and reports that were published between 2014 and present time about RPA on the web. Seventy two per cent of these reports/articles were case and survey-based studies on RPA implementation while 28 per cent were opinion pieces discussing the potential and impact of RPA.
We carefully studied these reports and articles and extracted key information about what so far has been achieved through RPA implementation as well as what is the state of the RPA market.
It was disappointing to discover that the vast majority of media coverage failed to report on key performance indicators such as costs saving (only 44 per cent reported their savings), return on investment (only 9 per cent), improvements in service in terms of speed (61 per cent) and accuracy (46 per cent). It seems that even though that RPA has been around for at least 4 years, firms are still shying away from revealing the benefits this technology solution offers.
Reported about speed improvements
Reported about accuracy performance
Reported about cost savings
Reported about Return on Investment (ROI)
Which RPA provider is getting media attention?
The relatively young but fast growing RPA market presents a one-time opportunity for a service provider to emerge as the ‘winner takes it all’.
Much of this depends on the reputation of the service provider as a market leader. Reputation is often developed via press releases and media coverage as signalling ability and quality to the market as well as word of mouth by satisfied (or unsatisfied) customers. Hence, how much media coverage service providers get makes a difference in establishing the service provider’s reputation in the market. Our study has found three front runners, Blue Prism, UiPath and Automation Anywhere, whose leadership position in the RPA market was signalled via significant media coverage.
Blue Prism was the most prominent with 34 per cent of the media coverage identified in this study. Automation Anywhere and UiPath followed closely behind, with 28 per cent and 26 per cent of media coverage. The remaining media coverage spread across numerous providers.
RPA provider discussed From media coverage perspective, it is clear that the game is still on and there has not been a “winner-takes-it-all” market leader. It is therefore imperative that these three players continue to invest in media coverage as a signalling strategy to attract clients and investors.
Who is buying RPA solutions?
While it is difficult to extrapolate from our dataset who is the big spender on RPA solutions, a clear indication from our sample is that the banking and financial services were widely reviewed by the media (37 per cent) with regard to their engagement with RPA projects.
This is not surprising in any way considering the high volume, labor intensive and repetitive work of some business processes in the banking sector, making it an ideal candidate for automation. But the media coverage has also paid attention to RPA activities carried out by the hi-tech industry (20 per cent) and professional services (16 per cent).
Only 20 per cent of our sample have reported significant job losses while 80 per cent reported minimal and medium job losses
In particular, media coverage of professional services is of interest as many of their reports offer insight into the impact of RPA however they fail to acknowledge the heterogeneous potential and effect RPA would have on different sectors.
For example, Deloitte, claims in its report that RPA has continued to meet and exceed client expectations across different aspects of their operations including 92 per cent for improved compliance, 90 per cent for improved quality and accuracy, 86 per cent improved productivity and 86 per cent improved cost reduction, however, offer little in attributing these benefits to a sector. Healthcare is another industry that received attention in the media (12 per cent), in particular with some interesting case studies of RPA implementation. For example, Mid Central District Health board created an up-to-date patient management system through RPAs 24x7 capabilities which resulted in saving amounting to up to 2.5 FTEs.
Media coverage by industry
Sectors that have very likely pursued RPA implementation however are lightly reported in the media are public sector, retail, manufacturing, construction and higher education. Considering that large scale, repetitive and rule-base processes in these sectors, it is surprising that no evidence of RPA implementation has surfaced so far through the press.
RPA: Does it really save you money?
One of the purported benefits of implementing RPA is its impact on cost savings. Unlike humans, software bots do not take vacations or coffee breaks and are not making mistakes, hence are expected to improve efficiency and significantly reduce operating costs.
For example, Synergy reported US$2.3million in annual costs saving following the implementation of RPA. In examining media reports, we considered four main categories of cost savings; slight improvement (up to 33 per cent in saving), medium improvement (34 per cent- 66 per cent), significant improvement (67 per cent- 100 per cent), and exceptional improvement (over 100 per cent). Considering the high expectations for costs saving, we were surprised to learn that 54 per cent of RPA implementation in the media reported slight or medium improvement in cost savings. Forty per cent reported significant improvements, with only 6 per cent that reported exceptional improvement following the implementation of RPA.
How much did cost saving Improve by?
Clearly, such findings suggest that the expected cost savings have not materialised for the RPA cases reported in the media.
Our research suggests that one possible explanation is that by implementing RPA, companies do not reduce headcount as well as do not redesign processes to improve end-to-end efficiencies. As a result, RPA solutions are standalone islands of efficiency that don’t improve the entire value chain. This is rather a concerning message to leading service providers that RPA solutions are not delivering the expected costs saving. In addition to costs savings, we sought information about the average return on investment (ROI) for RPA implementation projects. Despite the importance of ROI in IT projects, there is so far very little information in the media about what firms have managed to achieve. In our sample, only 9 per cent reported what their ROI was.
The impact on service speed and accuracy
Another of the purported benefits of RPA technology is its ability to outperform humans in terms of speed and accuracy. In terms of service speed improvements, our findings show that again, only 49 per cent of the cases reported significant to exceptional impact of RPA on operational speed while 51 per cent reported light to medium improvement in operational speed.
How much did Service Speed Improve by? In terms of process accuracy improvements, our findings show that 98 per cent reported significant accuracy improvements.
How much did Accuracy improve by? These findings indicate that so far the true value delivered by RPA has so far been in the form of service accuracy while both service speed and costs saving remain relatively moderate as compared with market expectations and ROI is pretty much unknown.
What has really happened with jobs?
Many analysts predicted the loss of millions of jobs following the introduction of automation. Our findings show that this is not the case.
Our dataset suggests that there were three job loss categories; minimal impact - up to 5 FTEs, medium impact 5 to 20 FTEs, and significant impact – over 20 FTEs. Our findings show that only 20 per cent of our sample have reported significant job losses while 80 per cent reported minimal and medium job losses.
Obviously, automation has not impacted jobs to the same extent previously predicated. One reason for this outcome is that many companies have retained employees to deal with exceptions, incidents that software bots cannot handle as well as job redesign and retraining has been common practice where software bots took over humans.
Impact on FTE Job Loss due to RPA
Of those who reported job losses, our study found that 56 per cent impacted job loss directly. However, the remaining 44 per cent of reported entries had no need for additional hiring. Instead of currently held positions being removed, anticipated new positions were the jobs lost instead.
Job Loss versus New Recruitment
Our findings (44 per cent - no need for additional hiring) also support an observation made by Horses for Sources (HfS) that RPA will enable companies to slow down recruitment and manage both ongoing and transformed operations with current staff.
In addition to predications around job losses, analysts have suggested that automation will also result in significant improvements in job satisfaction in the sense that work is likely to become more meaningful, offering new opportunities for upskilling and retraining.
Our findings show that companies have reported that there were either slight or moderate improvement in job satisfaction (79 per cent) and only 21 per cent reported significant of exceptional improvements. Overall, 61 per cent reported some effect on job quality because of RPA implementation.
How much did job satisfaction improve by?
Considering these findings, we conclude that RPA implementation has not resulted in extensive job losses and the expected benefits for employees have not fully materialised.
This study sheds light on how much progress has so far been achieved with RPA implementation based on media coverage of nearly 160 reports and cases.
While we have learned on the impressive growth of the RPA market in the last four years, our research shows that there are some weak spots in this market that both users and service providers must pay close attention to their negative impact on market development and the adoption of the technology.
First, without clear signaling of the ROI of the technology, potential adopters will remain hesitant and would treat RPA as a local solution to an isolated case of inefficiency rather than a technological component within a much bigger networked technological solution that is attempting to optimise a service value chain.
Second, this study shows that RPA has achieved only moderate improvement in costs saving, service speed and accuracy, while service providers have consistently promoted RPA on the grounds of exceptional improvements.
Why, then, there is such a tremendous gap between the expected and realised performance? These results call for (a) greater transparency on behalf service providers with regard to the realised improvements in service costs, speed and accuracy.
Using cross-industry cases can help service providers consider various scenarios of RPA implementations and model them to illustrate potential improvement subject to various conditions.
End-users need to insist on simulating the effect of RPA implementation prior to embarking on implementation as well as consider the impact on the service value chain. Just accepting providers’ word at face-value that RPA does magic is naïve considering the evidence we gained in the last 4 years of RPA implementation.
Last but not least, to increase confidence and signal ability, there has to be greater transparency and information sharing through media coverage and case studies by the three platform providers Blue Prism, UiPath and Automation Anywhere.
In doing so, the leading providers will be able to enhance their reputation by clearly signaling their areas of strength but perhaps more importantly finally differentiate their solution from the crowd.
Dr Ilan Oshri is a professor and Georgia Moore is a research scholar at the University of Auckland Graduate School of Management. Oshri is the author of 17 books on global sourcing and has extensively published on the future of work in both academic and managerial journals. This article is part of his research related to the Smart Sourcing and Intelligent Automation course he is teaching at the University of Auckland Business School Executive Education.
Get the latest on digital transformation: Sign up for CIO newsletters for regular updates on CIO news, career tips, views and events. Follow CIO New Zealand on Twitter:@cio_nz