If Charlie Chaplin’s character from the 1936 movie Modern Times lived today, he would probably be some sort of bored and frustrated data entry clerk. Instead of screwing in bolts with his wrench, he would probably click a fictional mouse compulsively.
After all, we perceive monotonous and dehumanizing factory work as a distant memory, blown away by the massive deployment of industrial robots. On the other hand, this kind of repetitive workflow has been transferred to the service sector. There, in the depths of their offices, millions of employees are engaged in tedious and mundane tasks such as data entry, invoicing, inventory management, and so on.
Fortunately, things are changing once again, thanks to the skyrocketing adoption of robotic process automation (RPA) across a wide range of different industries and use cases. According to Grand View Research's 2020 Robotic Process Automation market report, the global RPA market reached $1.4 billion in 2019 and is set to grow at a CAGR of 40.6% from 2020 to 2027.
The following optimistic forecasts are shared by Statista, which estimated the worldwide RPA market to hit $2.9 billion in 2020 and more than 10 billion by 2023:
The term “robotic process automation” can often be misleading. So, let’s clarify a pair of essential points.
First of all, it does not involve the use of mechanical robots but software robots, leveraged to automate repetitive and time-consuming business processes and streamline a company's workflows.
Second, this type of tool is often considered a branch of the wider AI realm, but it's something pretty different. Indeed, RPA can mimic human action while, unlike AI, it cannot do the same with human thinking.
For example, RPA can take care of updating customer data, generating periodic reports, or carrying out other straightforward actions. But it cannot analyze data sets, learn from them, and make decisions on its own like AI. This does not exclude the possibility of combining the two technologies to maximize business results, as we’ll see later.
However, such a difference is something to be aware of, as it implies certain limitations that need to be carefully considered when it comes to selecting the best RPA use cases for your company. That's why it may be worth dedicating part of the investment to robotic process automation consulting before actually integrating these technologies into your business.
Before delving into the limitations mentioned above and the best practices to address them, it can be helpful (and more motivating) to start with what robotic process automation can do very well.
The list of the most common RPA use cases is really long. To simplify our overview, we can split these use cases based on the sector in which they are effectively implemented.
This significant macro group includes the banking, financial services and insurance industries. According to the aforementioned Grand View Research’s report, the BFSI sector alone accounted for more than 29% of the global RPA market, ranking No. 1 among all major industries in 2019. This may be due to the ever-increasing volumes of data and transactions that financial services companies have faced in recent years.
Among the top RPA use cases in BFSI, we can count:
The healthcare segment of the global RPA market is the one that Grand View Research estimates will expand with the highest CAGR in the coming years. Actually, it’s not surprising considering the need to efficiently, quickly and securely manage a massive amount of private data.
Some of the noteworthy RPA use cases in the healthcare industry are:
According to Protiviti's 2019 Global RPA Survey, the telecommunications sector is among the most advanced in implementing RPA.
Some of the most relevant RPA use cases in this industry are:
The retail sector is not the most advanced in terms of RPA implementation. However, according to Protiviti's estimates, it dedicates about 0.11% of its overall revenues to RPA spending (double that of the manufacturing industry).
These investments target different fields of application, such as:
When it comes to automating operations in the manufacturing industry, precedence has traditionally been given to physical processes. Yet, non-physical processes such as gathering data about machinery performance may have greater automation potential thanks to RPA. This information provides technicians with useful insights to improve productivity as well as maintenance and worker safety.
But that’s not all. RPA in manufacturing can also be leveraged for:
Although RPA can act like humans (even faster and better) to perform a wide range of repetitive tasks, it cannot think like humans. Basically, it runs fast but it's still pretty dumb, just like Forrest Gump.
With these limitations in mind, we compiled a brief list of best practices for picking the right RPA use cases and prepare your business for implementation.
Traditional RPA lacks the analytical and cognitive abilities of humans or AI. Therefore, the best candidates for automation are rule-based processes, which follow well-defined patterns with relatively few exceptions.
On the other hand, RPA greatly accelerates repetitive and standardized processes. So it is worth implementing it for high-volume routine tasks, especially if they are mature, stable, and perfectly integrated into the company's workflow. After all, why invest in automating processes that could change overnight?
This is a fundamental and often overlooked requirement. According to Deloitte’s 2020 Automation with Intelligence report, only 38% of companies can rely on mature, well-defined and standardized processes. Unfortunately, immature and fragmented processes are one of the most important barriers to implementing RPA because they prevent a unified and synergistic workflow.
Last but not least, an important parameter of choice when selecting RPA use cases concerns their potential ROI. Processes with a high FTE are better suited to be automated because this will ensure greater savings. Also, consider the overall impact of these processes on your business workflow and performance and keep in mind their potential scalability to further enhance the benefits of RPA.
We have already mentioned that RPA can be merged with other technologies to overcome its limitations and maximize performance. As you might expect, the prime candidate for such collaboration is AI.
Based on Deloitte's 2019 Automation with Intelligence report, companies that have chosen this hybrid approach and implemented the so-called "smart RPA" have generally performed better, increasing their revenues by an average of 8.5%, compared to the 2.9% of companies that rely only on traditional RPA.
For example, artificial intelligence can complement RPA's skill of extracting data and compiling reports thanks to its analytical capabilities. In this way, data is not only collected but also interpreted to provide insights or make predictions.
Alternatively, the automated sending of emails and notifications on behalf of customer service and marketing departments can be combined with machine learning-based pattern recognition. This allows for greater personalization based on the analysis of customer data and preferences.
History teaches us that every great technological innovation, especially in the field of automation, involves great changes in the labor market. According to McKinsey’s estimates, more than 81% of physical work, 69% of data processing, and 64% of data collection activities could be automated with these technologies. This may lead to job losses.
The key to containing this painful side effect, while enhancing the benefits of RPA, lies in workforce reskilling. A good RPA adoption plan should consider retraining employees and redefining their job roles based on their future interaction with robots.
But it's not just about knowing how to interface with machines. The ultimate goal shall be to develop purely human cognitive skills such as problem-solving, creativity, and critical thinking. Something that, unlike data entry and other clerical activities, a machine cannot mimic.
Following this approach, hopefully, it will be possible to turn millions of workers into high-value professionals.
According to Stephen Hawking and a long list of other characters both real and fictional, "perfection just doesn't exist." Robotic process automation is no exception to this rule.
It’s certainly a multitasking tool that can be deployed for a myriad of different applications but comes with some drawbacks. It lacks analytical skills and struggles to interpret any event that deviates from the norm. Moreover, it risks generating significant social upheaval and being welcomed with great reluctance.
That's why any implementation strategy should focus on integrating RPA with AI, selecting the most suitable use cases, and above all investing in workforce retraining. Even though we're talking about machines, after all, it's still a matter of humans.