You read that right. Robotic Process Automation (RPA) in the present tense now offers capabilities that far exceed its early iterations of scripting, screen scraping and scheduled jobs. I know some others in the industry will disagree that RPA in its true sense is not an enterprise platform. I respectfully disagree. Several years ago, industry experts made the distinction between RPA and Intelligent Process Automation (IPA) which, in my view, only added to the confusion.
Currently the term “RPA” has much more name recognition than “IPA”. New and existing customers of the technology generally refer to “RPA” when in a literal sense it’s “IPA”. It is my thought that the concept of RPA is not static but is rather something that is evolving and will continue to improve over time.
As I have written before, one of the greatest boosts for the RPA space was coining the term “RPA” to describe what I and others had been building for years before it. Without something to call it, there really wasn’t an effective way to describe what the software program was doing. I have always looked at RPA as an entire suite of tools. Initially it was activities that clicked buttons and typed data into applications. After that Orchestration brought scale to the management and monitoring of the bots on Virtual Machines. Now we are in a stage where the next set of activities are mature, work reliability and can bring enormous outcomes and ROI within a small amount of time.
The promise of Process Mining as a Business Process Management tool for an enterprise is real and documented (Check out this demo with Claim Capital on Process Mining Claims). In addition, Task Mining can yield meaningful results and is imperative for the expansion or in some cases the start of an automation program. I will also say that traditional activities like OCR, Natural Language Processing, Object detection, and complex document understanding are all very reliable. In some cases, using the best in breed intelligent document understanding tool, like Indico Data, offers delivery on the promise of the technology with minimal upkeep and fast time to production. In addition, the Machine Learning model has gotten exponentially easier than it was just 1 or 2 years ago. The wide use of these models will become more prevalent and cost effective as the processing power of computers increase and the underlying data is cleansed and accessed more easily.
I realize that these tools are each in their own category and segment of the software space but to a bot they are all just another activity it can use to automate a task, which, in itself, is RPA.
Written by: Peter Camp, CTO & Founder