By Henri de Bruine, Managing Director at Decision Inc. (Pty) Ltd. UK
While Robotic Process Automation, or RPA for short, does not give you a big green button that fixes all your problems, it does remove a lot of headaches that hold a business back.
To put this into context, here are examples of some RPA use cases that show the processes that can be automated:
- Automatically login into a system as a user;
- Automatically navigate around the software, using menus and shortcut keys mimicking a user moving the mouse and clicking or typing on the keyboard;
- Reading values from a screen by interpreting labels. So, if the screen content or Web page is dynamic, the process will still be able to find the right value on the screen;
- Automatically reading values and data from PDF reports and putting them in data tables;
- Automatically inputting values into a capture screen by mimicking the user’s actions to do so. However, it can process it a hundred times over with different values for each pass;
- Reading emails and messages to understand their intent and the information they contain, for example is the sender placing an order, making a booking, looking for information, or sending a document such as an invoice or claim? Based on this intent, the system knows which process to trigger with the message;
- Scraping information posted and continuously updated on a Web page;
- Triggering a process in a system by logging in as a user and clicking a button or a series of buttons; and
- Evaluating many different pieces of information and deciding what the next best action or corrective action must be in the process.
But is this really something? Well, yes and no. Yes, because we are automating digital (in software) processes performed by information workers. No, because a small number of automation tools have existed for some time, but they have not targeted the enterprise space. Instead, they were generally used in the PC segment to automate clicking and typing tasks.
Originally, process automation was implemented to good effect in the manufacturing space for almost a hundred years. It started as mechanical automation of parts in the early 1920s. In the 1970s, huge manufacturing efficiencies were gained through the adoption of CIM (Computer Integrated Manufacturing).
Some might say that this digital RPA is doing the same as plant automation did in the 1920 to 1970 but is instead geared for the information worker. The truth is, it does more. CIM meant that controlled processes were configurable. But this did not mean that the process is necessarily more intelligent.
In its current form, RPA performs more than the mechanical replication of steps (and some decisions) in the process that it is automating. Artificial intelligence enables RPA to become more intelligent in dealing with requests like an email or a message sent by a customer. Natural Language Processing sees RPA interpret a message to understand what the context is and can therefore ‘decide’ who or which process needs to deal with it. Furthermore, Computer Vision allows RPA to analyse pictures and interpret their content. It can even process video streams to automatically action a call if unusual events occur on-camera.
So, for the technophiles that ask the obvious question: “Is this low-level system-to-system integration?” Well, no. Because it interacts with the user Interface of the system, it does not require the development of API-level system integration.
When it comes to business, RPA allows for two things to happen:
- Remove costs and errors from people performing tedious and repetitive processes.
This is similar to the impact manufacturing automation has had on the automotive and electronics industries. Things have become cheaper and more reliable when human constraints are removed from the process.
- Allow the business to do things it could not do before.
Imagine a situation where demand would spike a hundred-fold from one day to the next. It would be too costly and inefficient to keep a workforce on stand-by for the odd occasion when this happens. With digital workloads, the systems can access resources on-demand when the spike occurs and resume to normal operations when it subsides.
Customers expect instant response and action to be taken seeing a massive drive to transform the business into the digital future. But the constraints of legacy systems that make the transformation expensive and slow to take place can be overcome with RPA. RPA becomes the perfect middle ground of a tactical solution to digital transformation while the business churns through a three-year project to rip-and-replace systems. And with modern RPA, point solutions can be implemented in days, showing value very quickly.
RPA is an innovative set of tools and practices that allows business to migrate manual processes to automated digital processes. It enables the integration of legacy systems, online or external resources, and AI chat and email bots to move into the digital age without the massive rip-and-replace exercises of existing systems.
So, if you feel like you are not moving your business into the digital age quickly enough and systems are too complex and costly to integrate and automate, then you need to consider RPA as a quick way to move you forward.