RPA is based on a technology that emerged in the 1950s and 1960s and was developed to current standards by the industry's leading suppliers. This is the future? Think about AI.

Robotic process automation, the high-profile technology that automates routine digital tasks once performed by humans, has a long pedigree, dating back to the introduction of macros in the 1950s. The term itself and the basic technologies underlying today's products are of more recent origin. Here is an overview of the evolution of RPP.

First the concept, after the word

The current RPP tools have had many predecessors. Ted Shelton, founder and CEO of Robodomo, an RPA consulting firm, said that the basic concepts of RPA date back to the invention of macros in the 1950s, data scraping in the 1980s, and the advent of web scraping in the 1990s to index the web.

In the early 1980s, banks began writing data scraping applications to capture data from various financial data services such as Reuters and Quotron. Later, various vendors began building more sophisticated tools to cut and paste data from mainframe terminal applications to more modern web applications.

"The precise contours of what is called RPA are difficult to define as the term RPA itself only dates back to 2014 and has been used to name processes related to the automation of tasks in the tertiary sector, especially since automation itself as a principle has been the driving force behind IT for decades," says Raphaël Richard, an expert in artificial intelligence and founder of 24pm Academy, a start-up that offers artificial intelligence courses for executives. RPA's history can be seen to unfold in three major waves: The first wave included various types of programmatic automation techniques such as Excel macros. The second wave originated from test automation software developed by companies such as OpenSpan, which was later acquired by PegaSystems, the giant of BPM tools.

The third wave was launched by Blue Prism, one of the top three BPM vendors with Automation Anywhere and UiPath, with the development of enterprise-level automation tools with comprehensive controls and built-in security.

"At the beginning of the third wave, just having a vault and control room in good working order made a product stand out. Today, most solutions, but certainly not all, have the features required for the business," said Cottongim, adding that the third wave is coming to an end because the basic RPA engines have undergone very little change in recent years.

In parallel with the development of the RPA economy, CIOs are developing their automation macros or scripts with varying degrees of success.

The evolution of RPA is now entering what is considered a fourth wave, driven by the integration of so-called intelligent components such as OCR, machine learning and chatbots into RPA applications. Although these capabilities are not new in themselves, the maturity of RPA developers allows them to combine these capabilities with traditional RPA. Now that the industry has proven its ability to automate rule-based tasks, companies can start thinking about improving the workflow between robots and humans.

This allows some smaller players to position themselves in niche markets, such as the french online accountant (expert-comptable en ligne in Molière's language), which after re-imagining accounting processes by launching one of the first online accounting offerings, integrating image recognition systems based on deep learning or this sales forecast specialist.
Evolution of EPS: how it was renamed

Phil Fersht, founder and chief analyst of HFS Research, introduced the term "robotic automation" in a 2012 blog post about the dawn of a type of outsourcing that involved processes and technologies -- "with people as an option". Later that year, Pat Geary of Blue Prism (see below) added the word "process" to refer to a new category of automation designed to complement business process outsourcing (BPO) and business process management.

Since then, the term has become widely used, driven by the leading vendors that developed its core technologies - Automation Anywhere, Blue Prism and UiPath - and adopted by more than 150 different vendors that offer some flavour of BPO. Major enterprise software vendors such as Microsoft, SAP and Pegasystems have also entered the field. More recently, leading RPA vendors have begun to improve AI support and further integrate workflow platforms.
A Gateway to AI

RPA's rapid growth is a mark of its usefulness in automating many business processes. But Fersht pointed out that the $2.7 billion market for RPA tools is only a tiny fraction of the current $50 billion market for business process outsourcing, which relies heavily on low-cost human labor to perform the routine, repetitive tasks that RPA tools are designed to automate.

One of the factors limiting the growth of EPS has been the practice of the company adopting EPS as a tool, rather than building a centre of excellence to support it. As a result, only about 13% of companies have reached significant size. "Although there has been a lot of hype and hype, most companies are trying to do more than just pilot," said Mr. Fersht.

Fersht sees the RPA as a kind of gateway to the wider adoption of AI in business. He expects the RPA toolkit to evolve towards integrated automation platforms that will make it easier for business managers to combine automation, analysis and AI. These types of integrated platforms could help overcome some of the challenges associated with large-scale implementation of RPA-type capabilities.

The beauty of RPA is that it is the first time that many business users have had the opportunity to design processes using low code software. I think we will see RPA become part of a larger toolbox of solutions in terms of process orchestration and transformation.

The evolution of RPA, according to the top three RPA vendors

The vision of the UiPath leader

"Our journey began in 2005 in a tiny apartment in Bucharest, Romania. We had a graphical user interface (GUI) automation library -- similar to a screen-scratching software development kit for user interface (UI) automation -- and a team of 10 people. The tool allowed a user to double-click on any word on a screen and access a definition, such as a pop-up window or a Google search query. From 2005 to 2011, the software library was tested and used on tens of millions of computers by clients such as IBM, Microsoft and Siemens. It is this large installation base that has allowed us to refine our solution and make it work with any application in the Windows ecosystem. The tool was capable of intercepting clicks, scraping forms off the screen and sending the information to databases - that was the beginning of automation as we know it.

But the fundamental change came when a customer showed us how he was building these tools to train software to mimic basic tasks like data entry -- without the need for an engineer. We sent staff to visit the company, and then snatched a contract from Blue Prism, which had just coined the term "RPA" after automating back-office functions for banks. The client made it clear that this was the best use of our technology.

In 2012, our team saw the opportunity to develop a more business-friendly product, and the following year we effectively entered the RPA market. The first UiPath Desktop Automation product line based on the Windows Workflow Foundation was launched, which was perfect for process automation - making the solution fluid and easy to use. In 2013, UiPath experienced its first major RPA implementation when a global BPO used the solution to automate more than 50 processes for one of its largest IT customers".

The vision of Blue Prism

The origins of RPA date back to 2001, when Alastair Bathgate and David Moss of Blue Prism began to investigate how to create automation technology that could address operational inefficiencies in the banking industry where human workers perform repetitive tasks that span across corporate computer systems.

By 2007, Blue Prism had solved the long-standing problem of system interoperability by transforming the user interface into a machine interface and providing "code-free" connectivity to any system. This innovation allows digital workers to use and access the same IT systems and mechanisms as humans, so that they can automate processes on any system past, present and future.

Blue Prism has designed a robotic automation platform to perform tasks in the same way as humans, using an automated, pre-built and easy-to-control "digital worker". The term RPA was coined in 2012 by myself and marked a new category of software and a new market for this technology.

Another step forward has been to put RPA in the hands of professional users, which means that no programming skills are required to use this software. Blue Prism's Digital Workers are therefore pre-built, ready to use and no coding is required when using an intuitive central operating system to "draw, create and publish" automated processes".

The vision of the founder of Automation Anywhere (formerly Tethys Solutions)

"I dreamed of pursuing higher education after university. However, I quickly realized that I could have a greater impact by creating and accelerating technologies that are at the forefront of innovation".

I have spent my career in technology helping to establish new categories, and held leadership positions in the Internet, e-commerce and wireless markets before moving into intelligent automation.

Starting out in a number of different organizations, I saw first-hand how repetitive and labour-intensive daily tasks could affect human workers. Whether it's manually replicating the processing of a simple invoice or entering data from one system to another, I thought we could find a better way to use technology to eliminate many of these labour-intensive tasks while increasing productivity and efficiency, and freeing employees from their more mundane manual processes at the same time.

In 2003, at Tethys Solutions we developed robotic process automation technologies to help eliminate time-consuming tasks, although the term is not yet fully defined. ... Our slogan was "a robot on every desk". At the time, software robots were less sophisticated. In the beginning, most robots could automate individual tasks rather than end-to-end processes, and it was only later that artificial intelligence and machine learning capabilities were integrated.

What events have taken the APS industry to the next level? Four key events have occurred. First, the technology matured and was able to tackle larger scale and more critical deployments. Second, more millennia entered the workforce than ever before and had little interest in managing mundane and repetitive tasks in the workplace. Birth rates were also lower in developed countries such as Japan, forcing companies to adopt new ways of doing business.