The first time I heard the term RPA (Robotic Process Automation), I thought it was talking about actual robots moving actual things on a manufacturing line or warehouse. To be honest, I was a little disappointed when I found out it had to do with automating manual tasks and recording mouse movements and keystrokes. I want Rosie the Robot from The Jetsons!
Nevertheless, RPA has gained significant mindshare over the past few years and positioned itself to be the end-all-be-all solution for everything from data entry for a large, complicated ERP to auto-liking my Instagram posts.
Google Trends searches for RPA over the past 5 years
As with all new technologies, expectations rarely meet reality in the beginning. People get excited about what capabilities the new technology can deliver, only to have hopes dashed as the “true” use cases come to light. Gartner calls this experience of managing expectations vs. reality the “hype cycle,” and in 2017 they put RPA right at the top of the “peak of inflated expectations.”
Where do you think it is in 2019? Even if it’s no longer at the top, it’s well on its way into the “trough of disillusionment.” According to a survey from AIMultiple.com, at the end of 2018, greater than 40% of RPA projects failed to deliver on expectations.
Although this post may sound like I’m throwing shade at RPA, I’m not. I’m just trying to be realistic and manage expectations as we collectively use this new technology and get to know its pros and cons. I believe that, in the future, we’ll look back on RPA as Henry Ford’s manufacturing lines of the digital age. Especially as we start to layer in other technologies like machine learning (ML) and other artificial intelligence (AI) paradigms
Imagine this: you’ve got an RPA process pushing data into TrackVia, which is, in turn, performing some machine learning and reporting results, then managing the workflows and notifications when there are exceptions… now we’re cooking with oil! RPA + TrackVia + ML = a monster of a data processing engine and the ultimate solution for efficiency.
Or perhaps my thoughts are still in the “peak of inflated expectations?”
Inflated expectations or not, we’re excited to help our customers begin their journey with RPA in their organizations. One of our banking customers recently started exploring RPA after working with us to digitize a workflow that was previously processed by a small army of people. Now with RPA doing the data collection, more people can be re-assigned to higher value tasks.
Every company’s journey with RPA will be different. Here’s a rule of thumb as you look into your own RPA, this time coming from Craig Le Clair from Forrester. He calls this his “rule of five” to help determine the right processes to automate.
- No more than five decisions. More than five and you’ll need a rules engine.
- No more than five apps. RPA is naturally fragile; it will break if a software vendor changes their interface. Keeping the number of apps low reduces your exposure.
- No more than 500 clicks. Again, due to the nature of RPA and its inherent fragility, keeping clicks down makes RPA scripts much easier to manage.
So there are guidelines for moving forward. We need to know our limits in order to measure the true costs and benefits of this new technology.
We at TrackVia are huge fans of RPA because it’s part of a grander formula that supports our efforts to surface information from data. We are constantly trying to provide analysis and visibility into the information customers are storing in TrackVia so they can ultimately learn from it and make better decisions. As I said above, we believe RPA is the infrastructure and groundwork for not only automation but the vehicle for realizing the power of AI. These are pretty lofty visions. Hopefully, expectations won’t be too far from reality.