Robotic Process Automation (RPA) and Chicken Feather
Introduction
Conceived from robotic processes in manufacturing industries, RPA refers to automating repetitive tasks. Bots are the tools deployed to automate such processes. Then comes Intelligent Process Automation (IPA), many companies use both RPA and IPA synonymously. However, “intelligence” construes AI aspects such as machine and deep learning algorithms Nowadays, the term “RPA” is used for cases where we put in top-notch AI-based solutions and “IPA” is not that famous.
We also have something as IT Process Automation (ITPA) specific to IT Service Management (ITSM) of application support and maintenance realm. For example, ITPA involves automating incident management in a way that handles incoming alerts, analyzes, verifies, and prioritizes them, notifies the appropriate parties, and then, upon the desired action being taken, completes the workflow and closes the ticket. Again, some companies synonymously use IPA and ITPA.
We might ponder that we need more differentiation. RPA is automation for the end-user (surface) while ITPA is more behind-the-scenes automation. Both are designed to improve efficiency, boost productivity and cut costs, but the way they are ultimately used is slightly different. Also, RPA is more “reactive” where IPA or ITPA is more “proactive.” The detailed Discovery phase can help in identifying which solution fits the bill.
Major types
Unattended robots completely free the human employee from performing very tedious tasks
Attended robots take place on the employee’s desktop, where the attended robots work collaboratively with human employees, for processes that require human intervention
Pros and Cons
Bots provide better accuracy, improved AHT, low code, front end, low investment, and high ROI, cuts mundane form- and filed-filling activities for the agents — overall efficiency increase.
Multiple discrete systems (CRM, transaction system, and so on ) that agents handle to get data and input into another can be automated by RPA.
The flipside includes security and maintenance.
Then what about APIs and Chatbots?
The ideal scenario of RPA is the automated control of existing graphical user interfaces (UI) when the application does not offer an API. Cases like metered API usage (cost), lack of time and resources, and disparate legacy systems favor RPA over API.
Where does the chatbot fit in? Interactions that chatbot handle cannot be considered as repetitive or mechanical, it will definitely require some complex aspects of the AI world (NLU). Of course, we can configure less complex scripted bots that do not involve AI. The same applies to RPA: simple bots, bots with NLP capabilities, and bots to make decisions with AI rules engine (we are not there yet).
In reality, there are cases where all these are combined to give an RPA solution. A chatbot collects the invoice process request along with invoice id, an API helps fetch some fields from ERP or CRM, bot executes the series of processes, and details are emailed to the recipient.
Tools in the Market
Blue Prism, Ayehu, Automation Anywhere Tool, NICE, and UiPath are significant products that we can leverage to create RPA and IPA tools. Renowned strategy is built once and reuse as needed (we may come up with something like “Bot foundry”). The inherent challenge is how far we can reuse it as it depends on the specific tools, applications, environment, databases, and so on, involved. Of course, we can develop in-house tools with generic frameworks such as Python.
Approaches
In general, proper due diligence of the tasks that can be identified for RPA — tasks that are repetitive, rules-based, and manually intensive; those tasks, which classify as low complexity and high volume. Even tasks with high complexity and high volume can be good candidates for RPA alongside Optical Character Recognition (OCR) and Business Process Management (BPM).
In the previous section, you might have noticed NICE (who is specialized in the voice recording realm) is placed in the top players' list of RPA. Any guess why they have reached that place soon? First in place, they excel in recording, which puts them in a better position to analyze the transcripts and come up with automation solutions. Especially, their abilities in “Desktop Analytics” play a vital role in designing automation solutions.
Thus, it has much to do with analyzing the interactions and designing the specific bots for superior CX.
What the heck is “feather” doing here?
Which part of the chicken you think has the highest potential for proteins? The upper part, liver, skin, and so on? Unexpectedly, it is the feather that has abundant keratin. The problem is that this protein is in locked form (titbit: we have structural (like hair and nails) and functional proteins [like enzymes and hormones]). Researchers are using chemicals and microbes to unlock structural proteins and derive the pure form of proteins such as keratin supplements.
Similarly, if we can unlock the unstructured data in the recordings by analyzing the transcripts, the sky is the limit. Targeted automation solutions will rule the CX world. We will have the potential for zero direct support. Imagine 100% recording and transcripts running through neural networks, the scope for automation increases to multifold.
Conclusion
Just want to conclude that we should use quirky analogies to grab attention and register the intended message in the audience's minds. Like the “chicken feather” here. Nope, just kidding?
This is an attempt to sensitize and induce our thoughts on emerging trends of AI-based automation solutions set to change the way CX is delivered. With self-help and process automation, the scope of innovation and ideation is wide open for us to contribute. Believe something new sparks on your head on these lines as you complete this article.
