As Robotic Process Automation (RPA) adoption is rising with every passing day; most enterprises which have already implemented RPA and have been using it for a while now, have come to realize a few bitter truths:
In short, a human is indispensable in the RPA journey!
These points also bring us face to face with the fact that deterministic (or rule based) automation, which RPA caters to, can only cover a certain distance and raises a few questions about the future of automation, which are worth pondering.
RPA integration with AI and Intelligent automation with built-in self-learning capabilities is the industry’s forward looking answer to all these questions. This is popularly called as RPA 2.01. As per analysts, “It is expected that the RPA market is going to reach about USD 8.75 billion by the year 2024”1. But the path isn’t straight forward. Accuracy of intelligent and cognitive services still pose the biggest challenge! And to counter that challenge, “attended automation” undoubtedly has a pivotal role to play in realizing this RPA 2.0 vision. And to answer the question pertaining human involvement, yes human does have a role to play and we will uncover the details in this blog.
Let us first understand attended automation, look at how it addresses the current roadblocks in RPA and broaden the scope of process automation. We will then see its role in realizing the Intelligent automation vision.
Defining Attended Automation
Attended Automation is the flavor of assisted RPA where the bot sits on the agent machine and is triggered by the agent for automating tasks in a big process. It saves agents’ precious time automating swivel chair2 operations and mundane tasks, enabling them to do value added tasks which require human judgement and intervention. Our whitepaper on “Demystifying Assisted Automation” takes a deep dive into the topic based on the usage and context (e.g. in Contact Center).
How does attended automation address the roadblocks in RPA and broaden the scope of process automation?
For processes that cannot be automated end to end using RPA, we recommend some quick checks and balances:
Attended automation helps leverage the benefit of automation along with human intelligence where end to end automation is not possible.
Handling exceptions and fallouts in the unattended RPA processes:
For automated process which are error prone, it is a good idea to build them as Attended automation processes.
Role of Attended Automation in Intelligent automation journey
These days’ use cases that involve reading of scanned documents using OCR capabilities, are already in definite consideration set for automation. The use of the cognitive capabilities like image based automation, OCR, computer vision, AI, ML mark the start of intelligent automation.
A genuine challenge in the use of cognitive capabilities is the correctness or rather incorrectness of output from the cognitive services providing the capabilities.
In such a scenario, a “human in the loop”3 is a key capability to fill in this accuracy gap, make corrections to output where required and proceed further with automation.
To take an example of Invoices processing in the Accounts Payable flow where the invoices are received from vendors over an email. These invoices are read and details are fed into SAP after which they are taken up for further verification and processing. Often invoices are scanned and OCR is used for reading data from the Invoices, which gives the data a certain accuracy. Here attended automation can help to orchestrate this automated process and present the data read from the invoices by OCR for human eyeballing and correction. The corrections are fed back to the OCR model to train the same. This improves the accuracy of the OCR model needing less manual intervention over time.
Another example from Insurance could be claims processing where machine learning (ML) is used for getting the claim classification and in case of low confidence of the ML algorithm, it needs to be routed through a human for making decisions. Again attended automation becomes handy to manage the exceptions and based on the choices made by the human, process the claims further. The decisions are also captured and fed to the ML model for training.
RPA industry is now closest to bridging the gap between pure play RPA and AI/ML capabilities which were usually considered as distant capabilities.
What a teacher is to his students, attended automation is to intelligent automation. It is set to be a perfect teacher with the tool and means, to train the students (OCR, ML, Vision and the likes) and build a bright future to take automation to newer heights, seemingly unimaginable today.
So, if your Enterprise is on the digital journey and has started with RPA; without wasting any time, make sure to add attended automation to your toolset so that you are ready to embark the Intelligent Automation aka RPA 2.0 journey and reap the benefits from the same.