Companies both small and large are full of repetitive tasks that are performed by the employees. RPA (Robotic Process Automation) seeks to solve this by enabling software bots to go in and automate these tasks. Some tasks are trivial but you have to do them often thereby wasting time and driving you nuts (like copy-pasting values between two IT systems). Other tasks may be complicated as you have to do them rarely but they keep popping up a few times a year and bog you down for hours (how exactly was it that I setup a new customer in our system?).
In the old days most of the above challenges were solved through new features in your ERP (Enterprise Resource Planning) system or by doing business process design. With more and more companies moving towards distributed IT systems it has become much more common to apply RPA and focusing on creating good API’s (Application Programmer Interface) that can facilitate such process. For legacy systems without API’s or support it has often been possible to integrate them using simple screen clicking bots that are well-known from the world of software testing.
With the advent of AI and machine learning the range of tasks that RPA can solve has exploded. AI is not good at solving general problems but solving specific tasks (even ones requiring modest amounts of context) can be implemented efficiently and work just as robustly as a human. This differs from rule-based systems in that the actions performed may depend on multiple inputs and are driven by training examples instead of manually coded actions. This allows automation of processes such as redirecting a customer email request to the right department. It could however also be more complex processes requiring very specific expert knowledge like running through the lengthy process of establishing the credit worthiness of a customer.
The beauty of AI is that today you can use very different kinds of data as input. Making a bot that can understand what is in a picture, understand text, or even a phone call is no longer as difficult as it used to be. We’ve implemented numerous bots for highly specialized applications like counting people in surveillance pictures, appraisals based on text descriptions and other tasks which would normally otherwise have required following long repetitive work processes. So why not get started and figure out how you can use RPA to create value in your company today.