The digital age has brought us many beautiful things, but it has also brought digital drudgery.

By Rudolph Janse van Rensburg, chief of technology at TreasuryONE

Where we once filled out forms in triplicate, we now copy and paste the information with hundreds or thousands of mind-numbing, repetitive keystrokes. You probably are aware that physical robots are doing many of the repetitive and often dangerous physical jobs that humans previously performed. Industrial welding, deep-sea exploration, and disaster response are all areas where these physical robots are increasingly useful.

Now, there’s an equivalent in the digital world. It’s called RPA. RPA – Robotic Process Automation – is a way of automating processes so that repetitive and manual digital work is done by software robots, or “bots”.

Say you have employees who work in your customer service organization. As part of their job, they must create reports on any issues customers experienced in the 24 hours.

Previously, they would tediously copy data from the CRM system, paste it into the correct fields in a Word-based report template, convert it to a PDF, and then email it to their boss. Such a task could eat up hours of valuable time every week.

With RPA, it could be completely automated, freeing the employees up to do higher-value work – such as providing better, more personal assistance to customers. Moreover, the beauty of RPA is that once the bots are programmed correctly, they will never make mistakes. They are 100% accurate 100% of the time, never take sick days, and can work around the clock.

What can be automated?

Virtually any repetitive, technology-based task or process can be automated. From simple copy-and-paste commands to complex tasks–such as sensing when a customer is upset and escalating the issue to a human worker–many of your front- and back-office processes can be automated.

Basic commands, such as opening an application, signing in (inputting a username and password), cutting and pasting, calculating, and saving and closing a document can be automated. So can processes that currently require humans to interact with applications such as CRM, accounting, ERP, and supply chain systems as well as databases.

The gateway to digital transformation

Automation is at the top of CEOs minds these days. They see it delivering real benefits such as reduced errors, improved employee productivity, and cost savings. More importantly, they know it is essential for digital transformation.

In fact, in a recent research project on priorities in process and performance management, business research institute APQC found that 69% of the surveyed enterprises listed RPA as the foundation for their digital strategies. It’s no wonder that by 2022, the market for RPA software and services will top $4,3-billion.

Further, according to Deloitte, 53% of businesses have started their RPA journeys. The number will rise to 72% within the next 24 months. If this trend continues, RPA will reach near-universal acceptance within five years. The benefits of RPA adoption are significant, according to the Deloitte study. Payback was reported at less than 12 months, with an average 20% of full-time equivalent (FTE) capacity provided by robots.

Not surprisingly, 78% of those who have already implemented RPA expect to increase investment in RPA over the next three years significantly. But RPA benefits your employees, too, by making their work “more human.” What does this mean? With RPA, they now have more time to apply their human talents and capabilities to more interesting and higher value jobs.

By creating a workforce that is both human and digital – with the bots doing the rote, repetitive work, and humans responsible for the work that requires creativity, intuition, and judgment – workers are happier, more productive, and more likely to stick with their organisations, according to a University of London study.

A human/bot partnership

At its most fundamental level, automating a process with RPA means that you “record” the way a human works. Thus, when you create a bot, you are programming it to imitate all the keystrokes someone makes in the course of completing a particular task.

Although automation can massively improve the efficiency and accuracy of a company’s processes, it cannot do everything. There are times when bots need humans to make critical decisions. For example, when a chatbot detects an unhappy customer, or in a borderline mortgage application case, human intervention is warranted. Attended bots can do this. Attended bots automate processes but from time to time require human attention to complete a process.

Exception handling for better bot resiliency

Exception handling is a simple form of attended automation. When a bot hits an issue it cannot solve with its existing resources, the bot notifies the user prompting them for help on what to do next. For example, take document verification before making final approval for an insurance claim.

In industries like insurance, thousands of claims must be processed quickly and accurately to ensure internal compliance, keep costs in line, and maintain customer satisfaction at the highest possible levels. RPA is the perfect solution for processing these claims, as most of them are very similar.

However, at times additional input or approval is needed by an employee to complete a process. This might be a disputed claim, an upset customer, or possible fraud. Attended bots can be programmed to identify these instances and refer them to a human, to ensure proper processing.

A smarter way to automate

When RPA meets artificial intelligence (AI) new automation capabilities are created. Cognitive automation leverage bots capable of processing unstructured data thanks to AI and machine learning (ML).

What does this mean? Most of the data in the world is unstructured – it doesn’t come in standard, cookie-cut forms like a database with its orderly defined fields or a spreadsheet with its neat rows and columns. An employee is exposed to e-mails, chats, website content, Word sheets, and other data formats. How is a bot supposed to make sense of the various information and filter out the relevant data?

Ordinary bots are limited–they can only process structured data by repeating keystrokes that have been shown to them. Ask a standard bot to respond to a customer service email, and it will not know how to proceed. However, AIpowered bots can do much, much more. For example, a cognitive bot would be able to look at a utility bill and extract the relevant information out of it so it could be paid, because it knows how to recognize patterns in data that previously only humans could understand.


Just as in the Industrial Revolution of the late nineteenth century and the computing revolution of the mid-twentieth century, RPA is the revolution for the twenty-first century. All these revolutions drove business innovation and made our lives better and easier. And, not incidentally, we became more prosperous as well.

RPA bots take over repetitive, mundane tasks to free up knowledge workers to do more value-added work. A recent University of London study showed that workers in workplaces who had been augmented–not replaced–by bots scored 33% higher on rating their workplace as “human” than their non-bot-augmented counterparts.