What is behind the huge cost reduction potential from Robotic Process Automation (RPA)?
According to a report from Blue Prism, as of today and overall in the industry, more than 80 million jobs in the United States and 15 million jobs in the United Kingdom are good candidates for automation. Based on empirical evidence, several sources suggest that productivity increase can be as high as 80%.
Robotic Process Automation refers to a new technology category. It is a software robot that emulates the work previously done by people. To achieve this, the RPA software robot interacts with existing enterprise software through user interfaces, e.g. by interpretation of the screen signal.
Business users can create software robots that interact with existing enterprise systems just like humans. Currently, the RPA software is best suited to replace humans for so called ‘swivel chair’ processes. These are processes with media breaks, where humans take inputs from one set of systems (for example email), process those inputs using rules, and then enter the outputs into systems of record (e.g. ERP system).
Process automation through enterprise software changes is expensive
The primary reason why RPA was born is the missing automation capability, i.e. insufficient Application Programming Interfaces – APIs, and media breaks between enterprise software systems. An alternative is to achieve business process automation by building completely new systems, but this is often far too expensive. The result is a large and growing IT backlog of change requests from the business side.
During the last few years, several RPA software vendors have established themselves with an increasingly standardized offering. RPA software packages are designed to a large extent for non-programmers to use. The users can automate processes by following simple configuration steps or by recording key strokes and mouse clicks.
Software vendors also offer capabilities such as data analytics, machine learning and natural language generation to mimic human learning and decision making (IPA – Intelligent Process Automation). Once processes are fully automated, machine learning can also be used to analyze large amounts of process metadata (logs) and based on that unlock previously not visible optimization potential.
Achieving cost reduction, while maintaining local competency
Contrary to other cost reduction efforts, whereby expertise is often relocated to low cost areas, RPA helps to sustain local competencies. This can be very beneficial for high cost countries such as Switzerland. Lower level tasks may be taken over by robots, however, expertise and thereby also judgement and decision making remain local.
Robotic process automation accelerates processes, and it is easy to scale-up by adding more execution resources (robots). Meanwhile, experts can be redirected to focus on tasks that require judgmental skills and higher-order thinking. Together, these effects can lead to a massive productivity increase, while at the same time reducing operating costs.
Manual execution of business processes is error prone. A spelling error or wrong number introduced by a human operator can lead to process failures, which are expensive to correct and can adversely impact customer satisfaction. With the introduction of RPA, a significant quality improvement can be achieved. Increased automation leads to lower error rates and fewer process failures.
A further big advantage of RPA is the very short payback period. Robots operate existing applications independent of the underlying technology, so no expensive multi-year IT development projects are required to automate business processes. In comparison to that, RPA integration costs are very low.
A well-defined governance and control framework is required
RPA software does not easily fit into existing governance structures, because it can neither be classified as human capital nor IT system. It is therefore crucial that robotic process automation governance is clearly defined, and respective policies are created or updated before go-live of any RPA related projects. Among other aspects, the responsibility of the IT department with respect to RPA and how it is delineated from business responsibility needs to be discussed and agreed.
Often RPA based solutions are developed in an agile mode. A target business process is identified and automated to serve as a pilot or proof of concept. Such a pilot phase can be completed within a few weeks. As it takes more time, the adoption of an adequate control framework typically lags fast paced implementations of RPA solutions. In this context it is important to initiate work on the control framework as early as possible. The control function should work hand in hand with the RPA pilot project to understand specific risks and to define necessary checks and balances.
Especially when moving towards Intelligent Process Automation, which means that robots operate with more freedom, roles and responsibilities need to be carefully assessed and closely monitored. Where are the limits with regards to decisions or actions taken by a robot and how can such limits be established? These are important aspects to ensure a safe operation of RPA software.
Standardized software packages are available
Several software vendors have started to establish themselves in the fast-moving robotic process automation market. Examples of RPA software providers are Blue Prism, UIPath and Automation Anywhere. The most successful vendors have packages that are converging, with more and more similar features. Vendors are also evolving towards a similar approach for customer engagement through service partners. In the following sections, we describe the commonalities and distinctive features of the packages offered by the different RPA software vendors. A comprehensive vendor by vendor analysis is also available in a research paper from Forrester.
Integration with 3rd party components is rewarded
A major distinctive feature between the software vendors is their platform connectivity. Some vendors for example provide a proprietary solution for analytics. Other providers decided to focus on the core robotic process automation aspects of their software and instead provide interfaces to 3rd party components such as Google ML for analytics or Tableau for visualizations. The industry tends to reward an open approach towards 3rd party components. This can be seen by the increasing popularity of respective software vendors.
The robot market place is a newly emerging trend
A newly emerging trend is the digital workforce market place, where organizations can share their pre-built robots. In this way, it is possible to deploy robots in a fraction of time it takes to build them from ground up. This approach works particularly well for business processes related to widely used software packages, such as SAP or Salesforce. For example, there could be a pre-built robot to login to SAP and another pre-built robot to create a new SAP sales order. These are two standardized process steps that are part of many different business processes in the industry. Organizations can instantly download robots and deploy them as building blocks, to be used in their wider business process automation projects.
Training is provided at low cost
Software providers offer trainings and specializations on different levels. There are courses available for business users, solution consultants, RPA specialists, machine learning engineers, etc. In most of the cases, trainings are offered online at low cost or free of charge. Upon completion of trainings, some providers also offer exams and certifications are awarded to the successful participants. To deliver the trainings, many software providers have established dedicated partnerships with specialized firms.
Cloud usage vs on premise installations
RPA solutions have initially been deployed mostly on premise. This is because the first solutions focused on large enterprises with strong regulations on data security and privacy. With the RPA market maturing and overall adoption of cloud computing in the industry advancing, most RPA software providers have started to offer cloud capabilities as well.
Today, many organizations have already moved their applications to the cloud and are now running full processes in a more efficient manner. With the cloud, they achieved efficiency gains related to application support activities. However, organizations still need their employees to manually operate the applications. In such cases, cloud-based RPA software can make a difference by removing some or all that manual intervention.
Combination of RPA and advanced analytics capabilities
Whereas initially robotic process automation was focused on simple, straight forward processes, more complex processes can be automated with the help of advanced analytics capabilities. As most business processes have at least some complexity aspects, sooner than later, RPA tools that can support advanced analytics will replace tools that cannot.
In this regard, most vendors have started to incorporate capabilities such as text analytics to extract information from unstructured content into their software or natural language processing. Some providers also offer analytics to reduce maintenance of robots. To do this, some RPA software vendors have established partnerships with advanced analytics technology vendors. Other RPA software vendors mostly depend on their own algorithms in this area.
Enterprise grade security is a key consideration for process automation. Authentication, access control and end-to-end data protection are essential for RPA software packages. Logging, monitoring and reporting capabilities complement security provisions and are also broadly available.
How does a robotic process automation software package work? What types of components are offered? Again, we observe a convergence in the offering from the software providers. In general, there are three components that can be identified in most of the software packages.
Most of the software packages on the market offer a centralized control room or orchestrator. It is where RPA operators go to schedule, deploy, execute and monitor the digital workforce. The control room ensures a clear separation of automation from control aspects. It is usually available as a web-based interface and offers analytics and reporting capabilities that can be used to optimize the deployment efficiency of robots.
The Robot or Bot Runner, as it is also called at certain vendors, is a runtime environment where the robots can run in attended or unattended mode. Runtime environments are typically established on user desktops. However, it is also possible to install them on virtual machines, in data centers or on the cloud.
A development environment, also called studio or bot creator at other vendors, is a tool that can be used to model the robotic processes and then to deploy them. This is done visually, meaning without code (drag & drop), or by recording steps that are manually executed by a user. In terms of usability and level of technical background required for development, some vendors are more advanced than others.
Robotic process automation is an emerging field with significant momentum. Forrester estimates, that by 2021 overall in the economy, there will be more than 4 million robots doing office and administrative work as well as sales and related tasks. RPA is still young but has already gone through the initial phases of the technology life cycle. It is now proven to work, stable and ready to be deployed at an industrial scale. We are at the beginning of an exciting RPA journey.