Demystifying Robotic Process Automation - Rise of the Digital Workforce

A software company employing less than 1,000 professionals is expected to become the largest digital employer by 2020. Sounds insane?

That’s true. Automation Anywhere, one of the market leaders in robotic process automation is expected to deploy more than 3 million bots by 2020, that's the equivalent of around 7.5 million humans. [Link] (Assuming 1 bot provides efficiency equivalent of 2.5 humans)

That’s the scale automation, robotic process automation, in particular, is expected to achieve in the coming years. In the past few years, RPA or Robotic Process Automation has become increasingly popular. According to Gartner, RPA is the fastest-growing sub-segment of the global enterprise software market. RPA software revenue is expected to clock USD 1.3 billion in 2019, up from USD 846 million in 2018. Companies are increasingly coming up with requirements to deploy bots for multiple purposes - control expansion costs, reduce existing cost, reduce errors or improve customer experience. Each company may have its own reasons to explore RPA but this sudden interest in the field has made us wonder - whether this is really panacea to our problems? Or, maybe it is just another technology riding its hype cycle at present? In this six-article series on RPA, we have tried to decode the concept of RPA, understand its applications and market landscape, and provide an approach on drafting a roadmap for RPA for any organization.

What is RPA?

According to Gartner, "Robotic process automation (RPA) is a digital enablement technology that predominantly leverages a combination of user interface (UI) and surface-level features to create scripts that automate routine, predictable data transcription work.” RPA, robotic process automation, is essentially the automation of processes using robots. Two keywords to note here are “processes” and “robots.” RPA is unlike ‘Industrial Automation’ where robots refer to moving arms on conveyor belts or machines helping in loading and unloading of goods. In RPA, ‘processes’ refer to digital processes and ‘robots’ refer to digital bots. In other words, a bot in RPA is like a software which is programmed to carry out a process which was earlier carried out by a human. We will come to what kind of processes can be programmed in the later part of this post.

RPA is an application of technology to automate processes which can be broken down into steps using business logic. In other words, if a human being is carrying out a task which can be written down into step-by-step instructions, then it can be automated using RPA. ‘User Interface’ (UI) and ‘Surface-level Features’ mean that the automation workflows can be developed using front-end user interface with minimal coding requirements; while, surface-level features mean that the software uses screen-recording functionality to understand the process flow.

We can see the breakup of a process and scope of automation in Figure 1.1.

Figure 1.1: Manual Process Vs Automated process flow

How RPA is different from industrial or any other automation?

Automation has been there for decades. What is so different with Robotic Process Automation? Why has this become the new buzz word? In traditional or industrial automation, the focus of automation has always been on doing physical work - lifting up heavy blocks of goods, loading or unloading of vehicles with heavy metallic items or moving items from one conveyor belt to another. The robot designed in industrial automation can carry out only one task for which it has been designed (hardware as well as software).

In robotic process automation, the focus is on automating digital tasks - the tasks which are done on a computer or an application. The robots, or bots as they are called in RPA, can be reprogrammed to carry out multiple separate tasks and the entire designing happens on a software.

Now, the next question that pops into the mind: since the RPA is a system to system automation, how is it different from APIs? APIs are written to enable systems to talk to each other. How is RPA any better than API? (API refers to application programming interface - Wiki) The answer to this lies in the manner in which RPA works. RPA works by accessing the front-end of your system ("surface-level features” from Gartner definition); while, APIs work by accessing the back-end of your system. In order to push/pull data through APIs, you may need to make certain system changes at the backend so that systems are able to talk to each other. However, in the case of RPA, no such back-end changes are required. Moreover, it becomes difficult to integrate certain legacy systems due to their design and architecture; the same can be easily managed by RPA.

Intelligent Process Automation - RPA and Cognitive technologies

RPA bots are capable of automating tasks that involve semi-structured data, are based on clear cut rules and have zero learning capabilities. Complex scenarios where human intervention or decision making is required cannot be automated by RPA alone. The need for automation maturity that RPA requires to deal with complex tasks brings in the next wave of automation,‘Intelligent Automation’. (Refer Figure 1.2, Automation Maturity path)

IA (Intelligent Automation) is a layer of cognitive technologies over RPA that is capable of automating complex processes where human judgment is needed to make informed business decisions.

Cognitive automation mimics human behavior through vision, language, and pattern detection by using advanced technologies, such as natural language processing (NLP), machine learning, data and text mining.

Figure 1.2: Automation Maturity path

Architecture of RPA

In this section, we will not go into technicalities of the architecture of different RPA solutions available in the market. We will look at the architecture of RPA at a high level and understand the basic functionality of different components in the architecture.

Each of the RPA solutions has three major components (there could be exceptions, but majorly three components for setting up, running and managing bots).

For the sake of simplicity, we define our own names for these three components:

• Workflow Manager
• Bot Manager
• Control Room

If we were to talk about the top three market leaders in the segment, the following table could map our three components with components of each of the three providers.ig

Let’s understand the role of each of the components.

Workflow Manager is used for creating an automation workflow or a process which should be followed by a bot. This component is used to define a step-by-step guide to business logic which should be followed by the bot to carry out a task.

Bot Manager is a component where you create and map bots with the processes and workflows designed in the Workflow Manager. It may or may not be human-controlled based on the triggers that are set. Usually, this component has an interface with external applications from/to which data is to be read/written.

Control Room is more like an admin control which helps you control all the bots - deploy, schedule, monitor bots and processes. You can have a complete status of existing bots, active bots, inactive bots, process and other things in this component.

What are the benefits of RPA?

Organizations have found applications of RPA at multiple stages of the business cycle. Right from capturing lead data into the system to carry out the processing of the application to the final invoicing - companies are using RPA to streamline operations, reduce error rates and reduce the turnaround time for any kind of processing happening at the backend. RPA is used to process transactions, capture & process data, manage responses and communicate with another tech/ digital system.

RPA is deployed to streamline internal processes that lead to efficient deliveries. We can say that an automated process delivers faster & more accurately. We can broadly classify the benefits as follows:

1. Cost Efficiency
2. Accuracy
3. Scalability & Agility
4. Speed
5. Robust Compliance

In Figure 1.3 we have outlined each benefit along with a successful implementation success story.?

Figure 1.3: Benefits of process automation

In the next post, we will talk about the automation market evolution and the current vendor landscape.

Next Post: Understanding the Automation Market Evolution and RPA Vendor Landscape

This series of articles on RPA is being co-authored by Mitali Shekhawat and Vishal Bagla. Mitali Shekhawat is a Manager in Turnaround and Performance Improvement practice at Alvarez & Marsal.