Robots in the context of office processes and their relation to artificial intelligence
What does a robot mean in the context of the office, how is it configured and what is its relationship to AI?
Throughout history, the word robot has taken on different meanings in connection with office processes. It is only in the last few years that its definition has stabilized, and today it refers almost exclusively to automation in the context of office work using Robotic Process Automation (RPA).
It is a program that, like humans, controls other programs on the same computer. Its aim is to perform tasks that would otherwise have to be done by an employee. With RPA, the company’s processes do not need to be altered or adapted - they can be automated in the same way they are currently being performed by humans.
“Robotic Process Automation does not require extensive input analysis, changes in IT systems and applications, impact analysis, integration testing or coordination with other departments,” explains Jan Herman, who leads the RPA team at Trask. “It’s enough for the process to be described by the employee who performs it. Because the robot uses existing user interfaces, it can take over the process from the workers and start executing it in the same way they have been doing it, without affecting the function of other processes, applications or systems.”
According to Jan, the configuration of robots in most RPA platforms is done via user interfaces, in which the activities performed by the robot are built from predefined blocks with configurable parameters. “In theory, this means that non-programmers who have undergone specific professional training can also create programs for robots.”
How does RPA relate to Artificial Intelligence (AI)?
Many people equate RPA with AI. This may be due, for example, to the ambiguity of the term robot and the new RPA platform enhanced with machine learning. Historically, however, RPA has had nothing to do with AI. Originally, RPA did not include any components or concepts that could be considered AI, although it was often integrated with AI, creating remarkable projects in which the robot acquired data, AI evaluated it, and the robot then processed it further or inputted the evaluated data into systems and applications. The integration of these two separate technologies (RPA and AI) is still often used today to handle document extraction, such as invoices, purchase orders or delivery notes.
Over time, individual RPA platforms began to expand their product portfolio to include machine learning, chatbots, computer vision, and the like. Nonetheless, in essence it was still about integration, albeit within a single platform.
More recently, however, AI has become involved in new areas of RPA, not only to extend the capabilities of this technology to other areas, but also with the intention of improving the utility value of robots and increasing the stability of automated processes. Hence, AI is gradually becoming a part of RPA technology and is involved in the areas of process analysis, process understanding, automatic process configuration, evaluating the significance of deviations in application behaviour, and adapting to minor changes in the applications served.
“Although various use cases of AI can be found in modern RPA platforms to address diverse areas of RPA, RPA and AI should continue to be viewed as separate technologies. The domain of RPA remains application servicing and the domain of AI remains data evaluation,” Jan Herman explains.
Author
Jan Herman
Senior RPA Consultant / Team Leader
jherman@thetrask.com
