Robotic process automation (RPA) – A look at future technologies

Robotic process automation (RPA) is a new IT technology that more and more companies are using. It involves getting software robots (bots) to quickly and easily perform various routine tasks on the computer.

While some next-generation technologies get written about and discussed at length without having any real impact on everyday logistics work, other technologies are finding their way into many companies unnoticed. Robotic process automation (RPA) is one of these hidden technology champions.

RPA applies software robots (bots) to automate manual tasks performed on a computer. The bots use IT systems in the same way as humans—in other words, they work on the program interfaces that are usually navigated using keyboard and mouse controls. For example, when inputting data, the bot locates the relevant field on the application or web page interface and enters the corresponding data, which it retrieves in the same way from a predefined data source.

A typical RPA bot is neither flexible nor intelligent; it sticks rigidly to a predefined process. For example: “Open Excel file A in folder X. Go to cell B10 and copy the data. Open a specific website in the internet browser. Paste the data copied from the Excel spreadsheet into the first input field.” The bot can be taught this process quite quickly using the RPA tools available on the market—simply by performing the task once for it and setting a few logical links. There’s almost no need for programming knowledge. The forerunners of RPA bots were scripts and macros, which could also determine such processes. While these were usually limited to a specific application or website, RPA bots can work in any number of interfaces.

On the job 24 hours a day, 365 days a year

Typical tasks that RPA bots are currently already performing include completing input and form templates, opening and doing simple processing of e-mails, uploading and downloading files, compiling data, and performing calculations.

A bot can be used to automate those routine, rarely changing tasks and relieve people of activities that are often seen as monotonous and tedious. Bots complete tasks to the same quality and performance level 24 hours a day, 365 days a year. The only thing RPA bots can’t deal with is unexpected changes in the process; these cause errors that require human intervention.

In the future, bots will use machine learning algorithms to learn from errors in order to become more flexible. “Super-intelligent” bots like these are still a long way off. For the foreseeable future, artificial intelligence will not be able to respond quickly, appropriately, and reliably to unknown events in the way humans can. All that’s realistic is for bots to take some small steps towards intelligence, perhaps by using machine learning to process unstructured input data such as texts, speech, or images.