A Complete Guide to Robotic Process Automation

Business Process Services

June 18, 2018

In the past one year or so you might have read/heard about the coming age of ‘Automation’, which has the potential to take away many jobs, possibly yours too.

The umbrella of automation is very wide and is achieved as a result of elimination of human effort by a predictable response triggered by a technological substitute, typically a ‘Robot’. We have seen multiple applications of Robots in mundane/complex/unpalatable human tasks across industries. The applications usage widely ranges from the assembly lines of automotive giants to the Unmanned Aerial Vehicle (UAV) drones in the war fields.

While the former case of an automobile paint job on an assembly line can be programmed, and therefore automated, a UAV might need remote controlling from a human, and therefore cannot be classified as a perfect case of Automation. This is not to say that there are no perfect definitions, which classify reduction in human labor with the help of technology as Automation itself. In order to keep the scope of Automation as comprehensive as possible, let’s assume all instances of Robotic intervention which substitutes or reduces human labor as a case of Automation, for the purpose of the following discussion.

There is no Definition

In a cultural sense, we refer to a technological substitution of human labor as a case of Robotic Automation only when it visibly displaces human work, something like industrial robots do in manufacturing. And, therefore we do not suspect the case of Automation in a Soda vending machine or near-field communication (NFC) for entry in a metro station which also works on the principle of

Self-acting and Self-governing
Substitutes work done by humans
Uses technology to achieve the first two points

We can try hard to draw lines on what perfectly constitutes Automation but the application of Technology is too wide for one definition to perfectly fit in.

Robotic Process Automation (RPA): The Philosophy

Note: Physical robots have been around for ages and have been displacing physical/blue-collar human labor for quite some time now. For the purpose of restricting our discussion on RPA only, we will be discussing only automation of ‘Knowledge work’ and physical labor is out of scope for further discussion, unless otherwise stated.

RPA pertains to automation of low-end (back-office) knowledge work, primarily in IT sector, which in itself is ironic as some of these IT jobs, in India & Philippines, are considered to have displaced many jobs in the developed world, on the back of offshoring.

Karma? Not so much.

With time, industries advances and evolves to sustain the demand for increase in productivity. When Physical Automation freed up capacity in multiple industries, excess human labor was deployed in high-end activities resulting in increase in productivity, and therefore, affordability and accessibility for more customers. Think of excess bank employees who initially got displaced with the advent of Automatic Teller Machines, but eventually got deployed back into the banking system, while expanding the number of branch offices.

And so, the argument goes that we will see yet another round of creative destruction with displacement of low-end white collar workers moving up the value chain and subsequent higher productivity resulting in an even better world.

Would a call-center resource be freed to code the next Facebook? Many people are not entirely convinced. Are you? Optimization (globalization, outsourcing) induced inequality and job losses are real, as the developed world is realizing the hard way with the ascension of politicians promising to throw the baby out with the bathwater. Nevertheless, what does the future of Automation really hold?

As always the case with most contentions, the reality could be somewhere in between. A McKinsey study demonstrates the extent to which certain work activities can be automated, and not occupations. The approach talks about breaking down work undertaken in any occupation, at a unit activity level, and then predicts the extent of automation potential for the occupation based on automation potential of its constituent activities. This implies that though the potential of Automation is present in majority of jobs that we know off, the resultant human workforce displacement may not be as extreme. To put it in other words,’ RPA will automate some of the white spaces and tasks carried out by humans as part of many activities they perform, resulting in concentration of human efforts towards activities they add more value to.’

And, therefore, some people won’t lose jobs because Software Robots took them, but because the immediate productivity gain (30%-200%) are too high to accommodate as much workforce.

In a more normalized near-future, we can see Software-robots working along with humans and complementing each other’s work. The displaced workforce will join more robots, or not, to cater to businesses which were hitherto unserved. Productivity gains of RPA would be the driver expanding the size of the pie.

Robotic Process Automation (RPA): The Concept

Leslie Willcocks, Professor at LSE, describes RPA as a type of software that mimics the activity of a human being, in carrying out a task, within a process. Therefore, it can do repetitive stuff more quickly, accurately, and tirelessly than humans, freeing them to do other tasks requiring human strengths such as emotional intelligence, reasoning, judgment, and interaction with the customer.

There are four streams of RPA as follows,

Highly customized software that will work only with certain types of process only, say, accounting and finance.
General streams can be described in terms of a three-lane motorway:

The Slow lane is what we call screen scraping or web scraping
The Second lane is a self-development kit where a template is provided and specialist programmers design the robot
The Fast lane is enterprise software that can be scaled and is reusable

We can multi-skill each piece of software across these streams. RPA is lightweight in the sense that its deployment does not need a lot of IT involvement to get it up and running. Business-operations’ people can quickly learn how to configure and apply the robots.

Major RPA players

Firms like Blue Prism, Automation Anywhere, UI Path, Work Fusion and Pega are leading the RPA race to automate administrative back-office functions like Finance & Accounting, Admin, HR, Order Processing and Customer Service. Besides, any area of the business which has structured rules-based repetitive processes like the ones mentioned above could be automated with the available solutions.

HfS Research, an IT analyst firm, has mapped RPA players catering to specific business requirements as below.

Useful links: HfS Research has also ranked RPA capabilities at a vendor level in RPA Premier League. This Quora thread talks about implementation of RPA for a leading vendor in the space. You can read it here for more understanding on different kinds of RPA software packages.

Adopting RPA

Below are the 4 important considerations for adopting RPA,

Strategy: RPA could be tactically leveraged for cost savings. But it could also be deployed as a broader strategic tool, we can get a lot more out of it.
Launch: Getting the C-suite involved, appointing a really good project champion, and picking up the right process are important ingredients of ‘Launch’ stage. The process has to be stable, mature, optimized, rules-based, repetitive, and usually high volume. Organizations can start with a controlled experiment on visible bottlenecks or pain points.
Change Management: It entails persuading the organization to change and adopt automation.
Building a mature enterprise capability: Long-term users have built centers of excellence over time, usually within business operations, and have developed skills and capabilities within that center. They have people, who assess the feasibility of a proposal from a business unit. They also have the capability to configure a robot, install, and develop it, as well as the controllers to best plan how it fits with human work.

The Robot ahead

While the definition of Automation might span across multiple definitions, the purpose is not. *Shoshana Zuboff, Professor at Harvard Business School, proposes that organizations/managers deploys Information Technology (a super-set of RPA), to accomplish one or more of the below 3 interdependent operational objectives, i.e. to increase the

Continuity: Functional Integration, Enhanced Automaticity, Rapid Response
Control: Precision, Accuracy, Predictability, Consistency, Certainty
Comprehensibility: Visibility, Analysis, Synthesis of productive functions.

RPA, in its current form, makes a slam dunk on all 3 counts, and therefore is the technology in Business Process space today. And, while the technology is pretty much mature on its own, it will only evolve further from here, transforming itself into Cognitive led Automation.

So, where does it get dirty?

Businesses will erode, as customers take hitherto outsourced back-office work in-house or re-shore it. The service piece will shrink and be selectively applicable to limited streams of RPA deployment.

However, the biggest nightmare for the workforce would be the time when RPA expands into Artificial Intelligence (AI) space resulting into true Cognitive or Intelligent Automation. Just like Natural Language Processing or Face detection, immense compute capacity will power software to recognize patterns of problems, something referred to as Problem Management in IT industry, where the pattern recognition of a Knowledge Error Database (KeDB) leads to a standard solution, for all similar instances (incidents) in near future. This is a classic case of analysis driven knowledge-work done by a human until now, but could be the first amongst many complex activities to be displaced by Software Robots.

Automation + AI = Cognitive/Intelligent Automation

Once, we are past a threshold level of cognitive capability, compute power, infrastructure, data and cost-benefit case, we would have more to worry about, than just jobs!

For now, we can learn to apply automation for a better world.