Part 2: In RPA, not everything can be automated.

Bobby Patrick, the chief marketing officer at UiPath, said:

“Employees today struggle to perform under the weight of mind-numbing, repetitive work… As the study finds, the reality is that RPA allows employees to engage in richer interactions with others, perform work that requires more brain power, and make fewer mistakes.”

Robotic Process Automation uses rule-based software to perform business process activities at a high volume. Rule-based software means that for any RPA automation use case, we must have defined inputs and outputs. For instance, in the example in part 1; extracting data from a database to an Excel workbook; the data in the database is defined (array, JSON, integers) and the output is the Excel sheet.

The automation actions have to be the same every time and scalable across the business thus, creating a group of digital workers who implement the repetitive rule-based tasks.

While RPA improves the efficiency of a workforce, RPA does not work for ambiguous workflows; for instance, if the database changes or a row is added to the output Excel file, the bot will not run or return improper data. RPA is not yet good at handling variability or changes.

RPA bots are good at executing rules-based tasks i.e. X happens after Y and only if Z condition is met.

RPA is not suitable for processes that:

1. Require constant human oversight

2. Are too complex and have variances.

3. Deal with unstructured data

One of the most appreciated benefits of RPA tools is that they require no coding skills, which makes them brittle and unable to adapt to an organization’s changing requirements. Once a process changes, bots are rendered buggy.

My two cents: Your organization mustn’t fall into the trap of overestimating how many of its processes are suitable for RPA, and underestimating how much work is required to fine-tune process rules. As a digital worker automates most back-office work, areas of improving the process are materialized.

Here is a link to another article that might help paint a clearer picture — https://www.linkedin.com/pulse/5-things-you-need-know-rpa-robotic-process-automation-tony-walker-1f/

In part 3, we look at how we can move beyond task automation and expand their offerings to include intelligent automation (IA). This type of automation expands on RPA functionality by incorporating sub-disciplines of artificial intelligence, like machine learning, natural language processing, and computer vision.