Why Robotic Process Automation is »dumb« ─ but its use in SAP smart after all.
A training app conducts individual surveys via chat on a regular basis and provides the user with analyses for his training progress. A complaint is quickly categorized via the service chatbot of a large online marketplace and passed on to the responsible human employee if the bot cannot solve the problem directly.
Behind these rather commercial examples lies RPA technology (Robotic Process Automation): software robots that imitate human behavior and automatically process recurring tasks.
Information is often spread across several systems and is accessed and merged by the user via different applications. In this connection a software robot (bot) acts as an interface and can take a lot of tedious work off the user’s hands. RPA interacts with all necessary software systems simply via the user interface. This is why it does not usually require extensive implementation on servers. A prerequisite for the simple use of RPA is that the data is structured and undergoes a rule-based process.
Difference between attended and unattended automation
When software robots carry out recurring tasks, RPA technology is often subdivided into unattended automation and attended automation. One or another will also come across the term Robotic Desktop Automation (RDA) on the net which refers to the attended RPA - please also refer to our chart.
…is similar to a software assistant which is in dialogue with the user. Normally, a human activates the software robot to have a certain process performed automatically. If manual activities are required during the process or if problems (e.g. new scenarios) occur, attended RPA returns to the user.
A good example are assistants within the AFI solutions: they automatically compare data between different sources and make the user aware of discrepancies in orders or invoices for example.
…normally performs end-to-end processes and can therefore handle a process completely automatically without human intervention. However, such a software robot requires a certain infrastructure and maybe even process adaptions. Projects with unattended RPA are often more complex as they require additional programming effort.
Elimination of time-wasters: successful RPA performance in practice
Nevertheless, the investment in RPA technology is manageable and its benefits have been proven in practice.
First and foremost, the use of RPA saves a lot of time as SAP has also determined on the basis of customer surveys: employees gain between 15 and 30 percent more time for their core tasks.
This is also confirmed by a study conducted by Forrester Consulting on behalf of the RPA provider UiPath in 2018: two thirds of those surveyed even stated that their previous work had been restructured by RPA, creating more room for strategic work. 57 percent of the surveyed decision-makers also confirmed that the technology reduces manual errors. In addition, 86 percent noted greater efficiency among employees.
It is hardly surprising that many large companies focus on RPA. According to market research company Gartner, 85 percent of all large and very large companies are expected to use RPA by the end of 2022. The primary goal is to increase the degree of automation of business processes and focussing on the tedious, constantly repetitive tasks which are real time-wasters in daily work.
The RPA range of use is ideal for SAP processes
The RPA range of use is as diverse as structured processes are. RPA takes over the manual effort within unaltered, predictable business processes. SAP users in particular often have a high workload with recurring manual tasks and search for information from different applications before merging it back in SAP. Therefore, it immediately suggests itself to use RPA technology for processes around SAP.
For a better understanding of the range of use, we would like to provide you with a few functional examples. RPA technology can…
- check data,
- fill out forms,
- copy, paste and move data,
- perform calculations,
- read and process information from several different systems,
- read and process data from structured documents,
- create customer and supplier files,
- initiate orders,
- access websites and social media,
- maintain master data,
- execute if-then orders,
- enter orders
and much more.
A software robot can take over the determination of tax codes during invoice processing for example: if a tax code could not be determined in the SAP system, the bot performs this task using predefined logic, determines the tax code, enters it into the system and attempts to post the invoice. If the transaction does not operate, the bot sets a visible note/flag so that an employee can perform a manual check.
Or quite simply but very efficient: a software robot transfers information from an Excel document to an SAP purchase order free from defects and several times a day. It works easily, minimizes typos and reduces the workload.
The use of RPA opens many more doors and creates options to automate processes. The technology can be perfectly extended with artificial intelligence.
RPA can also be intelligent – but how?
RPA is not the same as Artificial intelligence (AI): unlike AI, RPA is not self-learning but does exactly what it is told. Therefore, the success rate for processing structured data is logically 100 percent.
For more complex processes, the use of RPA alone is not ideal. However, RPA and AI can be combined and create new synergy effects as a software team. The range of use of this intelligent RPA is thereby vastly expanded and also allows for the optimization of complex processes or the processing of unstructured data.
On the basis of document classification, the extension of RPA technology by AI capabilities can easily be explained: for example, the RPA bot always goes through the same classification steps using keywords such as »invoice number«. If there are additional languages or terms that the bot does not yet know, it can definitely learn them if an applicable AI function is integrated. As a consequence, the bot can directly apply newly learned languages or terms and routinely continue the same process, which in itself has not changed, with the freshly gained knowledge.
Why RPA is the opposite of process optimization
Although RPA technology can streamline processes as well as reduce errors and is in use 24/7, it does not improve business processes or IT systems. Comparable to medical diagnoses, RPA treats the symptoms but not the cause of a deficiency. Therefore, it is the goal of software robots to automate processes without changing existing systems. In »classic« process optimization, the focus is on fundamental optimization and process modification.
That is why a process analysis and proper process documentation, which also takes into account the people and the software systems applied, is crucial before using RPA. It is necessary to analyze exactly which steps are required to complete a task or how employees carry out the process to transfer it to the software robot accordingly. Precise instructions are required to display the logic of the process and to determine the sequence.
A predefined workflow is then set up through which the software robot always keeps going strictly the same way.
Conclusion: RPA is more than just a hype
Practice and the studies speak for themselves. Many of our customers are probably not even aware that they have been using the pioneers of this technology in their AFI solutions for years. But there is much more potential in this area - processes that might not even be on the radar of many customers can be accelerated and release unforeseen resources. It is a very exciting environment and we have done a bit of research for quite some time now. For example, we are currently introducing RPA on a large scale at one of our major pharmaceutical customers and thrilled when we experience the effects in practice which proves that RPA is more than just a hype.