Why you too should love Roger the Robot?

Heini Pensar Caption V2


I returned to OpusCapita a year ago after some time away. The decision to come home was easy. The billow of digital transformation was fast rising on the horizon, and OpusCapita was clearly on the crest of the wave. Working with robotic process automation, artificial intelligence and machine learning the past year has certainly been exciting.

The most interesting aspect about Robotic Process Automation (RPA) and quickly developing Intelligent Automation with machine learning and Artificial Intelligence (AI) is that they are creating a whole new automation culture in companies. At first they may seem too theoretical or technical to approach, but actually these technologies are quite easy to understand and relate to. If you are excited about spreadsheets – which admittedly most of us economists are – you should love robotics and machine learning, too!

Roger the Robot and his friends are already helping finance and accounting employees at OpusCapita and many of our customer companies: they enter product and pricing data into systems, validate and execute master data changes, process payroll data and sick leave notes, and even know when and how to issue a credit note to a customer. Roger is a software robot, a program operating in other programs. He is called a robot because it best describes the way he operates. The program monitors the activities of, for example, customer service specialists, and then copies them. He operates in a very human-like manner, which makes it easy to comprehend the huge impact he will have on our daily operations, and which is also the reason why we call him Roger.

Rule-based automation with RPA removes the need for robust integrations between systems and thus, heavy IT projects. Automation becomes quicker and more flexible simply because it works in a new way. The issue with Roger is that although he is extremely punctual and reliable in executing work tasks, he leaves the problems for me to solve. But imagine if Roger could learn from all the exceptions and errors he encounters in the processes. The next time he reaches a judgment point in the work flow, he could propose a solution and provide recommendations based on historic data and learning-by-doing.

Knowledge-based automation, often referred to as Intelligent Automation, takes the automation of knowledge work to this level. Artificial Intelligence (AI) as such is not a new invention, but until recently it had rarely been exploited in business services. The machine learning technology has evolved faster than anticipated over the past few years, and now it can easily process millions of rows of information in seconds to build complex rules and logic using neural network algorithms, for example. Coming from the BPO business I know the challenges of creating automation for supplier invoice coding. Performing the work requires understanding of context and language – not to mention the years of experience of an accountant. The machine learning program can overcome these challenges and combine both structured and unstructured data in order to form logic, an algorithm, as a basis for making its decisions. And it does it well: the suggested coding dimensions are correct 94% of the time.

What is great about these new technologies is that they allow business people to be in charge of running the development and deployment of new automation. RPA and AI can finally help to remove the manual pain points, which have been causing problems, and which ERPs, spreadsheets and macros have been unable to solve. They actually automate the tasks we thought were impossible to automate.

Published in OpusCapita Journal 2/2016. Read the whole magazine here