Now that robotic process automation has been a buzzword for a while, it’s descended from the Peak of Inflated Expectations into the Trough of Disillusionment for many.
Before achieving anything that is truly revolutionary, innovative technologies often cycle through a period of hype, a process Gartner branded and illustrated in its “hype cycle.” The cycle depicts a path that winds its way up to the “Peak of Inflated Expectations” followed by the inevitable freefall into a “Trough of Disillusionment.”
But eventually, the hype levels out and the genuinely game-changing innovations can start to reach their real potential via “The Slope of Enlightenment.”
Robotic Process Automation (RPA) is a good example of a technology suffering from the weight of the expectations that surround it – and a lot of misunderstanding about what it really can do. The answer is, a lot. But not everything. And it’s important to be realistic about what lies in between before you invest.
While the technology is a legitimate game changer, it doesn’t replace the need for data extraction or a document management system. There is a lot of confusion around the relationship between these parts of the process and we get a lot of questions: “Why can’t the bots do just it all?” Or, “Why do we need a bot at all?”
These are good questions. So in the spirit of helping you reach the enlightened slope a bit sooner, here’s an introduction to help you understand the where RPA can help, when to avoid it, and when to run for the hills screaming with your fingers in your ears.
What RPA Can Do
For structured and repeatable tasks, the “bots” of RPA can help you reap major productivity gains and virtually eliminate errors, with time savings you can leverage to save money and do more:
Um, yeah – Sunday, too
In fact, for document-related processing, capabilities like optical character recognition (OCR), document classification, invoice line item extraction, database look-ups, barcode recognition, and patch code recognition are all specific forms of RPA that are part of our ACE algorithm – you don’t need to recreate the wheel trying to replicate these with RPA.
Where RPA Fails
When RPA is applied to unstructured processes with variable data and/or “what if” scenarios, RPA is virtually useless. RPA can only replicate a 100% repeatable process. It can’t think. That is where machine learning (ML) and artificial intelligence (AI) comes in.
However, RPA can be used to either “feed” data by ML, AI or even other specialized RPA applications already built (e.g. OCR) by automating extraction or “being fed” structured data from these applications to automate portions of the workflow.
Separating the Help from the Hype
Don’t listen to anyone who tells you that RPA can replace workflow or business process automation (BPA). This is they kind of hype that can leave you hundreds of thousands of dollars poorer, just to find that RPA can only automate a tiny portion of your application – while potentially leaving you stuck with hundreds or thousands of hours of professional services on top of the software licensing costs.
As in the early days (read: still today) of enterprise resource planning (ERP) systems, such grandiose RPA installations are likely to fail in the early stages, abandoned and relegated to the dustbin of broken technology dreams. Run.
Get a Document Processing RPA Assessment
Let us help you identify if and where RPA can fit into your information workflows, especially for AP invoice processing, claims adjudication, employee onboarding, and credit processing.
We’ll assess what you need, identify potential solutions (whether we sell them or not) and set you on the path to automation Nirvana – or maybe just somewhere in the Caribbean as you celebrate your automation success with a well-deserved vacation.
Contact us to learn more about RPA and document processing automation
Chris Brown is the Director of Sales & Marketing at PSIGEN Software.