By Tolga Bayrak, Accenture Luxembourg.


Robotic Process Automation (RPA) is on the rise. My daily job reflects this. Ever since I joined Accenture in late 2016, I’ve worked on RPA projects. However, I’ve come to realize that a surprising number of people still have no idea what RPA really is.

I joined Accenture shortly before it was ranked #1 in the 2016 HfS RPA Premier League Table for Transformation Enablers. At that time, my knowledge of RPA was close to zero but my interest grew in parallel with the experience I accumulated. I am now fully convinced of the benefits of this technology and the impact it will have on future productivity.

This 7-step introduction to RPA is designed to dispel some urban myths. It’s the start of a journey in which I will share some of the challenges and benefits of RPA, best practices and success stories based on my experience at Accenture.  


There are no physical robots involved!

RPA is an entirely software based technology, so don’t expect to see tiny robots coming to your office anytime soon! The robots are installed on normal computers such as your laptop or workstation. At a larger scale, they can also be deployed in server environments and/or virtual machines (including in the Cloud) – making flexibility one of RPA’s greatest strengths.


RPA emulates human execution

The core thing a robot does is to emulate human execution. That is, it performs tasks on a computer like a human does: launching applications, surfing the web, copy/pasting information, filling in forms, you name it. In other terms, imagine one of those self-playing pianos, but for your computer.

As such, you can configure a robot to execute a typical business process, or parts of it. Of course, there are some limitations to what a robot can do, both technically and cognitively, which means that not everything can be automated. At Accenture, we have common suitability and eligibility criteria to tackle these questions.


Robots use regular user interfaces

The way a robot manages to emulate human execution is by using the regular user interfaces on your computer (note to non-techies: the “regular user interface” is the normal application windows you have on your PC, including the buttons, fields, etc. you use to interact with them). This is incredibly powerful because it means that you do not require privileged or backdoor access to an application to be able to interact with it. The robot can interact with almost any type of application such as web apps, Java, win32 or even mainframes (and there are far more lying around than you would expect).

The downside of this is that the execution speed of the robot is limited by the application because it is subject to the same loading times and latencies as regular human users. Nevertheless, a robot is still generally much faster than its human counterpart.


RPA is a non-invasive technology

You should think of RPA as an extra layer on top of your existing technology stack. A robot is like any other worker in your organization, carrying out tasks by using the same applications as you do. There is no need to rewrite legacy software or integrate RPA with business-critical systems, as they are most likely to be supported out of the box. As such, it is a non-invasive technology.

“A start to end delivery of RPA from analysis to deployment can be as short as two weeks”


There is no need for complex coding

When meeting with clients, I like to use the analogy of a student worker to describe how to configure a robot: assume he knows nothing about your job and that he has never heard of the software you use. These are the two main things you need to build a robot: (1) what are the logic and flow of the task, and (2) what are the applications needed to carry out the said task. Coding is object oriented, which means that you can define the use of a certain application once and reuse that object with other robots as many times as you want, significantly reducing complexity.


RPA can be rapidly deployed

Depending on the complexity of the process (number of steps, decisions and application screens) and organization maturity (existence of RPA objects library, strong Center of Excellence/governance model), a start to end delivery of RPA from analysis to deployment can be as short as two weeks. These short deployment times mean that Agile is particularly well suited to this technology.


Robots are managed by business users

Once approval has been reached to put a robot into production, it is transferred from the delivery team to the execution team. Typically, the latter is composed of business users who have a good understanding of what the robot does. Their job is mainly to manage the robot, make sure everything is running smoothly and take care of exceptions and incidents.


So what’s in it for you? Watch out for my next post.

Or feel free to contact us directly to discuss how Accenture can help you on your RPA journey. Because a picture is worth a thousand words and the best way to truly understand RPA in practice is by seeing it for yourself. We’d be delighted to show you!

Interested to know more?

Publié le 13 novembre 2017