Expert eyeTechnical

“For the Uninitiated” part 1, i.e. robotic process automation in a nutshell.

While process automation is no longer a new phenomenon, so in combination with the adjective “robotic” it still intrigues and leads to many questions. In today’s episode “For the Uninitiated”, Promity expert, Karol Pik, will present the problems of Robotic Process

Karol Pik – Robotic Process Automation Expert at Promity

Automation (RPA). The thing will be about software robots and virtual employees, as well as about the tools and applications of RPA-class systems.

People have already got used to the concept and phenomenon of process automation, but this adjective “robotic” still thrills some people. What exactly is robotic process automation?

Karol Pik: I would suggest the following definition: “RPA is a technology that automates repetitive and routine processes and activities by a robot performing man’s work in applications.

This definition partly answers the next question, namely the key difference between the process automation programs used for years, such as ERP class systems, and RPA technology.

Karol Pik: ERP class systems were and are intended to automate enterprise processes: large, complex, concerning the entire company, cross-sectional. These are e.g. purchasing, production and management processes. RPA, on the other hand, targets individual links in the process chain. And it is no longer about their automation, but about robotization, i.e. replacing the work done by a man with the work done by a robot.

Can we illustrate this with a simple example?

Karol Pik: A great analogy is the automation or ordering of a car’s production process by introducing a long time ago a production line at Ford’s factories. This is process automation. Robotization, on the other hand, is about replacing man, in this case with a machine, e.g. by introducing welding robots assembling car bodies.

What is a software robot? To many people, the robotization of business processes can bring to mind the humanoid cyber creature next desk, sharing office space with us.

Karol Pik: The name of RPA does not fully reflect what it should. This is certainly influenced by English, which does not always reflect what the author meant in a simple translation. A software robot is simply a software that does the work instead of a human, not for a human. Comparing this, for example, to a text editor: a text editor gives the user the ability to create text documents, while the robot would write such text instead of us.

Why do enterprises need RPA?

Karol Pik: The range of benefits is huge. The mechanism of simultaneous cost reduction and increase of effectiveness, quite rare in business, works here. RPA gives efficiency, accuracy, and repeatability of work. For human employees it can be tedious, routine, and thus burdened with the risk of simple errors.

Which workstations and processes are automated most often nowadays? When does RPA make sense? Which sectors have the most implementations?

Karol Pik: Ideal for robotization are processes that are boring, tedious, repetitive, time-consuming, and at the same time require focus and attention. In such cases, human error is common, and a robot does not make mistakes. As for implementation sectors, RPA has no preferred industries or areas of activity. Robotization of processes will work everywhere where we deal with a repeatedly performed activity that is standardized, which requires interaction with an application, and other kinds of system integration, for example by system calls, are difficult to implement.

Is RPA a temporary trend or a lasting change?

Karol Pik: It is difficult to talk about a trend or fashion in terms of RPA. This is probably just another step in technological evolution. The first attempts to robotize processes took place a long time ago. Today, tools used to robotize are more effective, easier to use, and are thus becoming more and more common. A smartphone is simply the next stage of telephony development, which appeared already in the nineteenth century. Will robotization oust humans, other applications? Probably not, just as the smartphone did not replace car navigation, despite the fact it is trying very hard…

At what scale does the implementation of RPA solutions begin to profit? Does such implementation require extraordinary expenses?

Karol Pik: This is probably another myth about RPA. This comparison comes to mind: from what moment is it profitable to have a car? Is there any universal answer? No, there is not. It all depends on the adopted criteria. If we have to perform an activity that we do manually, every day, and it takes us so much time that we are already thinking about it, then maybe it is worth considering the robotization of the process. It’s like with this car … You can go from place A to place B buying a bus ticket, being dropped off by a friend, or buying a car for several hundred thousand zlotys. You will achieve the same result, but in a different way and with different expenditures. It is just an issue of economic calculation and selection of tools. This car can cost tens of thousands, several hundred thousand or even several million zlotys. We have exactly the same choice when it comes to RPA tools. One can effectively robotize processes using free open source tools (e.g. Talend, Robotframework) or decide to buy a commercial solution. Here, the only limit is the amount of money one is ready to spend.

And side effects, such as unemployment growth? Maybe even disappearance of some professions? Is it not true that the mainstream perception of this technology is nevertheless branded with fear of losing a job and the social consequences of mass unemployment?

Karol Pik: Indeed, the topic is controversial for many, because of the stereotype: robots will take our jobs from us! That is why it is important to skillfully convey information about RPA and to familiarize with the unknown. Employees should understand that, by implementing RPA, the company is taking the most unwelcoming and disliked processes off their shoulders. It must be remembered that fear of progress and inventions has always accompanied us. After all, the same kind of arguments were risen when the steam engine was invented when the railway was established, and then when the first computer was invented. Of course, it is possible that the demand for performing “simple and repetitive” tasks will fall, but there will certainly be new areas, for instance, related to the operation of robots or the preparation of processes that will create new, more interesting jobs.

Well, but if someone only performs such simple, repetitive activities, then the “virtual colleague” will take them out of the equation.

Karol Pik: Well, “a colleague” … A robot is that kind of a colleague who will be happy to do what we do not want to do, and since we have already buddied up with “him”, let’s think about how to put such  inexhaustible willingness to work to good use. If “he” can develop, then certainly can we. Although, of course, resistance to change, including professional conversion, is inherent in human nature.

What are the main challenges – technical and organizational for the implementation of RPA? How to prepare for such implementation?

Karol Pik: Probably the biggest challenge is “sorting out” the company. By this I mean the order in processes, having their actual, current descriptions, having knowledge about these processes (what they use, what are the current and expected parameters). Another challenge is the organizational acceptance of robotization in the enterprise, seeking advocates and convincing decision-makers that the first robotic process will not shoot the company into a new level of revenues yet. Robotization is in itself a process, it takes time, people and understanding. It is necessary here to make a progressive adaptation of the company to the fact that there are “new virtual colleagues” who can work 24 hours a day, 7 days a week without asking for a raise… In technical terms, robotization does not have high system requirements. Currently available, popular desktop PCs can form a base to start the adventure with RPA.

Can the process of creating robotic process automation also be automated?

Karol Pik: This is a really cool question. Surely yes and it will probably happen in a while. The processes with standardized business rules will probably come first. Certainly, the development of artificial intelligence will have a big impact on the robotization of robots. However, there will always be a niche of specific solutions that will require dedicated or customized operations.

If we are talking about artificial intelligence, then we are talking about second generation RPA. Can these software robots learn?

Karol Pik: Yes, robots can learn. There is a branch of robotization, using artificial intelligence (AI) which “tracks” the actions of employees and builds models of behavior. At the end of the day, a model is created, a certain guideline according to which a properly constructed software begins to perform activities previously performed by a man. So it becomes a robot.

And what is the procedure for RPA implementation? Where to start? How is such a software robot created?

Karol Pik: A lot depends on how developed certain processes in a given organization are. Does it have business processes modeled (i.e. written and drawn)? Is there a list of selected processes (ready for the robotization) prepared? Are there any existing experiences connected to RPA? The first step is to choose a process or processes. If a company does not already have such processes modeled, then the processes are modeled by us. Here we use different techniques depending on the customer’s preferences (SIPOC, BPMN, Value Stream). Next, the process is analyzed and optimized. At this stage, we indicate places for possible robotization or automation. If our customer has no preferences related to the technology of robotization, we carry out an analysis of available technologies, which ends with choosing the most suitable option. We often precede this with PoC (proof of concept) in various technologies. Only at this point we create the robot which, after testing, can start working at the client’s company. It was supposed to be “in a nutshell”, so let us end it at this point. Everyone interested in exploring the subject of RPA and improving their organization is welcome to contact our experts.

The next episode of “For the Uninitiated” will be about the level of intelligence of artificial intelligence. We encourage you to read!