RPA is only a tool to perform automation. How to make this technology beneficial for business

Starting with a principal customers' question — what is a difference between RPA and automation? RPA is no way better than developers. But there are some issues that companies of all sizes deal with and which make this technology of a high demand. The first issue is a need of integration with legacy systems, IT-products without API and websites. RPA is a tool for very specific tasks. This technology allows developers to use a ready-made product to automate processes through a chain of combined resources. The second issue is a simplicity and availability of RPA products. Anybody with a knowledge of variables, loops and 'what/if' can use RPA-products. There are companies where backlog of tasks is scheduled for year ahead and new urgent cases occur each week and in this case they have to hire skilled developers to stay competitive. Then RPA as being a fast and available technology fulfill tasks and at the same time saving automation resources.

Applicability is a next question that arises around RPA. What we should consider here is that RPA is a specific tool for certain tasks, we cannot use it just for the sake of robotization and because it is affordable and easy to adopt. These benefits are quite elusive. An essential point is that RPA is used to connect several systems, but not for robotization within one system. Thus, it is more convenient to perform processing in SAP/1Ci using IT-products than RPA. A key factor here is stability, because RPA relies on certain algorithm and the smallest changes in a program interface may lead to robot's failure. This situation has nothing with processing and unloading performed with regular means. RPA is mostly made for automation of routine manual operations, for example, frequent data transfer from source to system, consolidation of data taken from different sources, document data recognition and distribution. Our customers often ask to automate the process of quarterly reporting, but this is not a suitable candidate for automation, because it happens quite rarely and will not pay off.

In general, we should think twice before approaching to emerging technologies. Trying to follow the digital hype may lead to disappointment. And RPA is not an exception. Our attitude to it has nothing in common with reality. Lack of skills and customer support are among problems before and after the process of RPA implementation. You can always ask IT-integrator for a help, but in this case implementation expenses will be higher.

Companies come up with the idea of RPA implementation differently, but there is only one way to do this successfully.

First, what we should do is to choose a right process for future pilot project. There are many options: to talk to an RPA vendor, who will definitely help in analyzing business-processes, to hire business consultants, IT-staffer will easily figure out the technology and where to apply it. But the most productive way is to listen to employees' complaints about loads of monotonous, uninteresting tasks, because ultimately their manual work will be the candidate to be replaced by RPA.

The second step is a pilot implementation where we are going to learn things the hard way and where specific for RPA processes of access negotiation, security and detailed algorithms become clear. It is not necessarily that pilot project or Proof of Concept will be profitable. Positive outcome of robotization is often a matter of scale effect. It is up to you whether to embed RPA on your own or to hire integrator. Here we have the same pros and cons that do not depend on the product like differences in competence level between staffers and hired specialists, deadlines, expenses, etc.

After successful implementation of the pilot project, we should think about RPA deployment across an organization. Deployment not only includes choosing and prioritizing the candidates for automation but also much important things, such as regulations concerning robots adjustments (they are highly depending on the system changes in which they work), employees' performance in a new digital environment and their interactions with robots, further development of the skills within the company or with the help of contractors.

As mentioned above, choosing a right platform is a principal point. It seems that every platform has common features, but you should pay attention to proven products of well-known brands with long-time customers. I would like to notice that it is not a matter of price, because prices on leading RPA platforms are at a one range. Here we should not consider foreign ones, because of VAT and exchange rates. It is less risky to use a proven platform, as ongoing projects make companies more responsible in terms of support and maintenance of their product and its further development. This ensures that company will not quit the market and leaving you with the necessity to pick out a new platform which mostly carries a risk of additional expenses and time to transfer already automated processes to a new platform.

The second important point is available information about a platform from active users. Ultimately, every company comes up with the idea of establishing an RPA Center of Competence. Thus, multiple of available information resources and support will help you to train new employees. Quality and service desk responsiveness plays an essential role here too.

The third point is a platform's functionality. Actually, RPA platforms tackle the same tasks but sometimes differently. The truth here is that every platform eventually shifts to convenience and simplicity regarding connection with certain software, web-sites or tends to embed additional tools, making product easier and more available for regular user. As an example, in our company we create simple neural networks in 3 clicks. These networks allow to save time significantly dealing with tasks related to classification or forecasting, without involving experts in this field.