It’s been a long time since I first played the game “Summer Games” on the C64. Admittedly, even then, the C64 was already getting on a bit, but the relatively uncoordinated alternating hammering on the two red buttons of the joystick is something that has remained in my memory to this day. All you wanted do was to be faster on the pixelated screen than your friend in the 100m run. Later, I used the first programs that simulated a mouse click and accelerated my progress in browser games – manual clicks were a thing of the past, at least partially. Even though many companies are not yet ready for it, today we are now at least talking about the partial automation of processes by e.g. Robotic Process Automation (RPA). But what is the idea behind this? Where do the challenges lie and what does all this have to do with Artificial Intelligence (AI)? – Let’s take a look.
First of all, RPA and AI are two closely related terms that will have a major influence on finance and the auditing of it in the future. While RPA’s approach is more process-oriented, AI is strongly data-driven. But RPA is also not a new technology like e.g. Process Mining. It has been spreading for 20 years and is mainly used for the automation of standardized and rule-based activities.
But what does that mean exactly?
What is a classic application?
Probably the best example is something that is likely to be known to everyone. You have an application open, such as SAP, and need to copy information to Excel to create a report. While you may be able to put up with doing this 10 or 20 times, copying and pasting data 100 times will start to get annoying. If you then have to switch between different transactions too, it will have taken a long time to create a report and the process will have been totally inefficient. This is exactly where RPA comes in. Tools such as UiPath, Automation Anywhere from IBM or blueprism can be used to define rules according to which the respective applications can be controlled.
But RPA also has its limits when it comes to unstructured data, for instance. Unstructured data, for example, is information on a scanned invoice. This does not make RPA useless, however. Together with other technologies, it can still provide immediate added value.
Artificial intelligence, or AI for short, is often misleadingly propagated as a magic wand for solving a multitude of problems. And indeed, it combines many technologies, such as data mining and machine learning, as well as speech and image recognition and semantic text analysis. For example, payment service providers such as Paypal nowadays use AI methods to differentiate between legitimate and suspected money laundering transactions. But the strength of AI lies in the analysis of data. However, AI cannot operate applications according to fixed rules, or read text from scanned documents, for example. For this reason, Optical Character Recognition (OCR), AI and RPA form a strong combination of advanced technologies that will be of great importance in the future.
However, a 2017 study by PwC revealed that only 5% of companies have reached maturity for AI and 15% for RPA. This is partly due to the high costs and partly to the lack of widespread know-how in these technologies. Companies are therefore faced with the challenge of counteracting this and preparing themselves for the future. For example, PwC estimates that approximately 45% of work can be automated, saving $2 trillion in global labor costs. This alone shows how great the potential for technologies such as RPA and AI is. The immense increase in productivity is something that not need even be mentioned here.
AI and RPA will have a big impact on finance and auditing! While nowadays recurring tasks are often done by junior employees, in the future, a (software) robot will take care of them. Employees will then have to use their expertise more specifically to make decisions that AI cannot make due to its nature. For internal audit, this means that the verification of controls will be automated by RPA and AI. The capacities that are already often too low today can then be used for value-creating activities. Full verification of data in the most diverse areas is thus a prospect that is getting closer and closer.
But why wait when you can already have zap audit for SAP today? 😉
This article is based on the following scientific paper:
Gotthardt, M., Koivulaakso, D., Paksoy, O., Saramo, C., Eds. Martikainen M. and Lehner OM. (2019). Current State and Challenges in the Implementation of Robotic Process Automation and Artificial Intelligence in Accounting and Auditing. ACRN Oxford Journal of Finance and Risk Perspectives, 8(2019) Special Issue Digital Accounting, 31-46
Practice-oriented sources on the subject of RPA: