In 5-10 years, companies designed as completely autonomous may appear. The majority of their operational processes will be done automatically and a human will have supervisory and managerial roles – says Andrzej Sobczak in a conversation with Maciej Chojnowski
Maciej Chojnowski: Let us start with organizing the terminology. Automation, robotization, robotic process automation (RPA), low code, no code. How do they differ?
Professor Andrzej Sobczak*: The most general term is automation which is the superseding of human work by machines. But as I will explain – it does not have to be a physical device at all.
Robotization is, in turn, the next step of automation where human labor is replaced by a particular device. In this case – a robot.
Automation and robotization are most commonly applied in the industrial sector. 5-7 years ago there was an increasing amount of talk about automation and robotization of the service sector. And that is where these terms start to take on a bit of a different meaning.
What kind?
When we talk about the automation of business processes then it is about the replacement of human by an information system. And when we talk about the robotization of processes then it really is a metaphor. We usually do not deal with a robot in a physical sense but with a kind of a software.
There are two other terms here: low code and no code platforms. It is a software with which you build applications but not in a classical way, by programming them, but by using ready-made components. We do it with the ready-to-use blocks which we link with the help of a graphic user interface (no code) or with a minimal use of programming abilities (low code). Thanks to it the software is created faster, it potentially has fewer flaws and is easier to implement. These tools are used to build applications which are also used for automation.
So we have low code and no code on one side and software robots on the other?
Yes. But in order to build these robots, a proper software is needed. This is where we advance to RPA class tools which allow for the creation of such a software robot. It works on the basis of a virtual helper: practically one of them can do things which needed human participation until now.
At the DARWIN 19 seminar you said that an increasingly important factor in the implementation of RPA tools is the streamlining of processes and the opportunity to design innovations, not just cost-cutting. Tangible benefits stemming from downsizing really don’t play a major role here?
It depends where. In Poland, there is a dominant belief that the main aim behind the introduction of robotization is the lowering of costs. But in Western Europe or the United States this is not the only success criterion. Over there, the companies see robotization as an opportunity to boost innovation and a fast introduction of new business products onto the market.
How does it work?
IT departments in many companies are experiencing implementation blockages. These departments have been, in simple terms, swamped with current work and big projects in the area of digital transformation. We also should not forget about the technological debt. IT departments are simply not reacting as fast as the business world expects them to.
The clients expect a flexible and easily implementable business offer. They also want to be surprised by new solutions. It pertains to the majority of service industries, including in Poland – banking, insurance, telecommunication – but also the public sector. An increasing number of people expect the state and city to provide them with modern services. This all boils down to implementing appropriate software solutions and the rebuilding of the processes. The robotization of these processes allows the implementation of such innovations faster, it gives new options. It will come to Poland in a while.
How does business depend on IT and why is it worth to limit that dependency?
Historically speaking, particularly in the Polish context, 20-25 years ago IT departments were at the forefront of innovation. They were the first stage of implementing the project approach, managing processing and the agile project management. Currently, in many cases – IT departments are a bit stale. Of course, they are trying to make progress with the resources, finances and time they have but business expectations – more, faster and cheaper – are really difficult to reconcile with the conditions in which IT functions.
Here emerges a particularly interesting concept for sales, marketing and operational departments – IT democratization, particularly in relations to automation and robotization. It is based on the fact that business departments – independently or in close cooperation with IT departments – create IT solutions for their own purposes. Thanks to it, there is a shortening of the development process and the solutions are better tailored to their needs.
How do IT departments see these changes?
In various ways. Some see it as a way to lighten the load of the simplest tasks. These solutions which the business builds itself are not that complex. It is not a central banking system but a supplementary solution, referring to the current needs in a company.
IT departments, on the other hand, are more skeptical. They remember the problems with earlier generations of tools used for democratization or they know the business will create something but maintenance will be on their side anyway. This is really about balancing the interests of both sides. Business needs to understand the concerns of IT and collaborate with them to tackle these concerns. IT, on the other hand, needs to understand business needs, as the most important thing now is the market position and organizing certain things can be done later.
There is no turning back from this because currently competitive advantage is not built on the process optimization and cost-cutting but on innovations in product, processes and marketing.
What kind of load does RPA take away from IT departments?
In the case of tools for the robotization of business processes, consulting firms and providers promote a model in which business implements them on its own. But, in a longer perspective, it is not a proper approach because then it is difficult to rescale this type of implementation and ensure operational support of the robotized processes. So this collaboration of business and IT has to be there. But it has to be a partnership. My understanding of democratization here is the participation of business in application development, robotization or automation.
Where is the main benefit here? For me, one of the basic obstacles of the current way IT solutions are created is the very beginning, i.e. doing the analysis what business actually needs. Business does not always think algorithmically, it does not always have its priorities straight. And even when they are organized, they evolve really fast because reality, the market or legal conditions change. Thanks to automation and robotization business can built certain solutions and test them fast. It does not have to concentrate on passing this knowledge to IT department. The IT department can expand on it later, of course, but the very beginning is much faster and natural for business.
An additional benefit is that a business working on these types of tools starts to learn algorithmic thinking on its own, and then it is much easier to collaborate with IT departments.
And what are the greatest challenges in the implementation of such systems?
The challenges can be divided into a couple of categories: organizational, technological and mental. When it comes to organizational challenges then the basic problem is that Polish companies rely on hints from consultants and suppliers and use robots for non-optimized processes. The first effects are then visible really fast but there is a big risk we will robotize a disorganized process, solidifying chaos. This is why one of the good practices is preparing the processes for robotization. So, centralization first, then standardization, and finally – optimization.
Another problem: when jobs done by humans are replaced by jobs done by a software robot then the knowledge of this process fades away easily. As a result, this knowledge has to be codified somewhere, for example – as guidelines. We can also adopt a rule that a particular process is ninety percent operated by a robot and only in 5 to 10 percent by a human.
Technological questions are another matter, particularly linked to security. It is a quite common concern that robots can do something unauthorized, not expected of them. This is why it is so important that, when introducing robotization, to be in close contact with security department. Certain standards as to how the robot is supposed to work, are set then.
Yet another matter is robot maintenance. It is something forgotten, especially at the beginning. It is relatively easy to build a robot but we forget that when systems, processes or regulations change, then this robot has to be modified, adjusted. Its software has to be up-to-date. There is a strong correlation with IT department here which is usually responsible for the maintenance of infrastructure on which these robots work.
Finally, the third challenge – the mentality. A risk exists that robots will be badly received, particularly in organizations where employees have less awareness. Robotization is seen as a kind of terminator, something which will come and destroy their world. It is not entirely true. An efficient communication strategy on management level is the key here.
What should an organization wanting to introduce robotization start with? What is the key to success?
Currently, the tools have a range of limitations. First and foremost, they are not intelligent. They operate on the basis of an algorithm that was designed by a human and in the case of an extraordinary situation happening they usually do not know how to handle it. And the costs of building robots are not small. This, in turn, translates into the fact that not in every organization introducing robots has economic justification. That is why the starting point ought to be the checking if, with the current generation of RPA tools, a given company has the necessary potential for robotization.
There are industries where robotization is currently doing very well. These are: the banking, insurance, telecommunications sector. Processes take place there which are relatively easy to standardize and optimize.
In other companies, on the other hand, given the scale, complexity, number of processes and amount of processed data, the cost of assembly, implementation and upkeep of the current generation of robots will be inadequate to gained economic benefits.
The evaluation of existing IT systems is another matter. It can turn out that we have such systems which are either difficult to robotize or it is enough to introduce small changes in the existing infrastructure and the robots will not be needed. Robotization is not a universal tool. It works in specific conditions and it ought to rather be the last attempt.
But if we have a company in which it is worth to do it then conduct a brief introduction, confirm it in practice and then build a robotic center in the organization because in the long run automation, robotization and also the artificial intelligence elements will be the source of competitive advantage.
What will it be based on?
If we look at the global economy, the biggest threat for companies across many industries, particularly the service sector, is not the competition that currently exist but technological companies that expand and enter the market which has been the domain of traditional players until now. This is the case in banking and it can soon be the case in telecommunication and insurance. And without treating automation, robotization and artificial intelligence as key competences, these companies can have big problems soon.
Automation is still connected with the need of standardization so everything which is automatically done, goes smoothly. Artificial intelligence, on the other hand, has to solve problems in a non-standard way. How does it currently look like and what can we expect in the next few years?
RPA tools present on the market today are the first generation. Even though these tools service standard, repetitive processes, the onset of partial artificial intelligence usage can already be seen.
For example, a classic robot will not know what to do with submitted invoices where the invoice comes in paper form and to be understood it needs to be converted into a digital form and processed. But if it uses an Optical Character Recognition system and then cognitive services which will process the OCR results and give a certain result which can then be processed further then we definitely have a contact point between robotization and artificial intelligence.
Main global suppliers of RPA systems are financed by investment funds. We are talking about hundreds of millions of dollars in the course of 2-3 years. And these leaders are clearly signaling that a major portion of investors money will be used on the development of cognitive functionalities of these tools. And this means that in the course of 2-4 years we can expect intelligent platforms for robotization.
Won’t these trends cause humans to be pushed out of the organization? Won’t automation and robotization be used as cost-cutting tools eventually? Or maybe in these automatized organizations a place can be found for several dozen people from management, as well as mid-level and low-level employees?
In this context I apply the idea of three horizons. The first horizon is a matter of one to two years The second: two to five years. The third: five years upward.
If we look at the first horizon, one to two years, then with the current generation of RPA tools, workers in Poland have nothing to be concerned about. Firstly, they will not be entirely eliminated from the process implementation. These tools will constantly require interaction with a human operator, be it in the form of supervision, servicing exceptions, extraordinary situations: further education, writing, maintaining changes.
In the perspective of two to five years no dramatic changes will likely take place, even though an increasing number of processes will be fully automated and the human role – minimized. Hybrid work environments will be created where software robots, built with the use of RPA tools, will operate.
Let’s remember about chat bots, voice bots, task bots which are a completely separate automation category of RPA tools. Here, the role of a manager is particularly interesting. It changes completely in this environment. Currently, he manages a team of fifty employees and a couple of not very smart robots. In three years, he is going to have fifteen employees and fifteen very smart robots. But this perspective is not that threatening yet.
I wonder how the job market is going to look in more than five years from now, when RPA systems develop significantly and achieve a high level of autonomy when it comes to decision-making. They will handle unusual situations perfectly. And it may turn out that, in many cases, a human will be eliminated.
What is more, In 5-10 years, companies designed as completely autonomous may appear. The majority of their operational processes will be done automatically and a human will have supervisory and managerial roles. In some enterprises we will have a brand new reality.
How will employees find themselves in this?
It is a mantra being repeated that introducing automation, robotization, artificial intelligence on a wide scale will cause a net increase in jobs. Maybe the old jobs are going to disappear and new ones will be created. I only have a concern that current employees, as a result of a competency gap, will not fit their new roles. And suddenly there can be vacancies.
There is a fight for programmers and data science specialists these days because there is more of them than the people who are ready to do the job. In 5-10 years, certain advanced competences will be expected even more. Will it be easier to find an employee then? No, because a mass of people will not be able to fulfill those requirements.
What can be done in this situation, then?
A never-ending cycle of education will be needed. For me, the democratization of automation is the first sign of an upcoming revolution in which business employees will be slowly able to prepare for a changing role and there will be more work with data or algorithm writing. Depending on the competences.
Will the need for human contact increase in an automatized world?
For now, automatic tools make mistakes but they also learn all the time. In a few years, you may not realize that you are serviced by artificial intelligence. Unless the law will enforce the disclosure you are talking to a bot.
On the other hand, where is a human being in all this? I can refer to a speech given by a CEO of a big international bank. He said that in ten years getting serviced by a human will be a privilege only the rich will be able to afford. Others will be serviced by bots and automatic systems. And, sadly, this seems to be the direction in which we are headed.
Automation and robotization is an interesting nugget of information today but it will become a standard. And in the future, human empathy will be something you need to pay for extra.
Not a happy conclusion.
This is the reality, I suppose. From the global perspective, in many cases, it will all boil down to assessment of economic risks and benefits.
*Andrzej Sobczak, Professor at Warsaw School of Economics, the Chief of IT Management Centre at Warsaw School of Economics. Executive Editor of Polish Forum of Robonomics. For a few years now, he has been involved in strategic IT management and corporate architecture (Enterprise Architecture) – including the application of this concept in the transformation of organizational operation – on the business level as well as IT. Presently, his science and consulting interests revolve around advanced automation and robotization of business processes. He is particularly interested in the role of democratization of data analytics and the robotization of processes in the implementation of innovations and improving digital skills.