Robotic Process Automation is one of the great misnomers of technology. Despite the name, there are no ‘robots’ involved – not in the conventional sense, anyway.
RPA (as it is generally known) is all about software ‘bots’ taking control of high volume, low-value repetitive computing tasks that people would previously have had to perform – data entry, invoice processing, mass emailing, basic service, and support functions etc.
But this is where the second misleading part of the name comes in. When we think of ‘robotics’, we tend to think of autonomous machines run by AI. But RPA doesn’t count as artificial intelligence. RPA ‘bots’ are really rule-based algorithms that are limited to structured data and defined inputs and outputs.
This is fine for automating simple, clearly defined, single processes. But it doesn’t offer any of the more sophisticated functionality we associate with AI – the ability to learn independently, for example, or to understand context, work with unstructured data or manipulate multiple disparate systems at once, the way people do.
Yet things are changing. Tech-savvy businesses have cottoned on to the fact that the value they can get from RPA is multiplied many times over when it is combined with other technologies – not just AI and Machine Learning, but the likes of advanced analytics and business process management (BPM).
This trend goes by different names – Gartner has referred to it as Hyperautomation, Forrester calls it Digital Process Automation, IDC prefers Intelligent Automation. But whatever the label, the meaning is the same – a way of thinking about process automation in a much more holistic, joined-up, dynamic way that pushes the realms of possibility way beyond what traditional RPA solutions alone could achieve.
Here are five ways ‘intelligent’ technologies are reshaping process automation.
Even in sectors like banking and finance where RPA adoption has been particularly high, it is estimated that as much as 80% of repetitive, low value IT tasks remain manual.
This arises from two issues – the number of tasks that involve unstructured data, such as filling in digital forms from unformatted documents, and the fact that many tasks remain ‘hidden’ as far as the possibilities of automation are concerned.
Intelligent automation has solutions for both. The Natural Language Processing (NLP) family of AI technologies are designed to work with data in ‘natural’ written and spoken forms, meaning they can extract information from sources inaccessible to standard rule-based algorithms.
Machine Learning can also be used to analyse user behaviour using software and identify barely perceptible ‘tasks’ that could be automated to streamline and speed up processes, such as switching between different applications on a desktop.
This ‘discovery’ phase of intelligent automation drags back the starting point of the automation journey a considerable distance. Hyperautomation not only allows more processes to be automated, it makes the decisions on what can be automated, leading to automation itself becoming a complete end-to-end cycle.
We might refer to the above as Machine Learning ‘automating the automation’. One of its effects is that it leads to a much more joined-up approach. The tasks taken on by traditional RPA are not only pretty narrow in their scope, they also tend to operate in silos. Through advanced analysis and the cognitive capabilities of Machine Learning, AI can by contrast link automated processes together into long mutually beneficial chains.
That’s when the impact of automation really starts to multiply. It also means that organisations have the power to scale up automation very rapidly and to high levels, because the work involved in deciding where to apply automation is machine-led.
Oversight of digital processes
The ‘discovery’ aspect of intelligent automation also has another benefit – it provides businesses with in-depth, end-to-end insight into all digital processes. As AI-powered systems analyse IT operations and usage, they create what is known as a ‘digital twin of the organisation’, or DTO.
As a virtual representation of all digital workflows, a DTO provides real-time intelligence and oversight that, according to Gartner, allows “organizations to visualize how functions, processes and key performance indicators interact to drive value.”
It is often argued that increased automation poses a threat to human labour. But the purpose of intelligent process automation is to support and augment human use of IT systems, driving efficiency and allowing people to focus on higher-value tasks.
Forrester has forecast that the way intelligent automation will shift the relationship between people and technology in the workplace could create $134bn in labour value by 2022.
As IDC puts it, the more that intelligent automation can replace human oversight of the technical side of digital business processes, the more opportunities people will have “to release their creative intelligence, and kick-start the man–machine collaborative learning cycle.” That will include using insights gathered from DTO to look at ways of improving and adding value to processes where the ‘human touch’ remains vital, such as customer experience.
Process automation for the remote workforce
Finally, intelligent automation also has a solution for one of the biggest trends affecting workforce organisation right now – remote working.
To date, RPA hasn’t been available for any GUI’s using desktop virtualization – which is exactly what most people do when they log onto work systems from home via the cloud. This is because remote desktops project a virtualised ‘image’ that conventional RPA bots cannot read.
But with computer vision AI, that barrier is removed. Computer vision allows an RPA algorithm to process what is input on a page optically, rather than relying on the structured data input of standard user interfaces.
This is potentially a big step towards exerting the kind of control and consistency businesses are looking for over dispersed workforces, as well as driving improvements in efficiency and security.