When is RPA not RPA?

RPA is primarily a term given to an age-old screen scraping technology that provided automation of existing computer desktop applications screens that users would have to navigate manually (keyboard and mouse). Be that a task (looking up a customer), a bunch of tasks (look up customer, validate for overcharges, send an email) or a process (pay an invoice, onboard a customer, check for fraud, reconcile and account(s) etc.,). A human may or may not be involved in assisting the RPA bots and various terms over the last few years were added (Unattended RPA, Attended RPA , Robotic Desktop Automation (RDA), Human in the Loop, Digital worker, Digital Workforce et al).

Ok, that's a fairly basic explanation but net net, that's probably the shortest summary you'll ever read. These automations can run on someones desktop or on a virtual machine locked in a server room. Some great Robot Managers and Control Towers were built to securely manage the deployments and running status of the "robots" regardless of which mode they run in. Throw in some nice IDE's, marketplaces / certification exams and a big market is born!

The lack of scale, increasing bot breakages and wrapping poor processes is a constant issue (as per what killed screen scraping 20 years ago) and I think we can all agree too, it doesn't matter how many bots are deployed but ultimately, how many hours are automated, errors are reduced, rework eliminated and response times back to the customer are improved. These are ALL great benefits from RPA when you can get them.

What drove the RPA movement was simply the "selling to business", who are often desperate for operational improvements, the "promise" of the ease in which you can build and maintain this "stuff". Some of the RPA IDE's are pretty slick and make everything look like a piece of cake (though this ease of use in itself, according to analysts, seems to be at the heart of the problem of the growth of broken bots but that's another story). The CTO isn't so easily fooled!

But what I'm asking here today is when did everything else in IT / automation / integration become RPA? If the RPA vendors IDE can integrate with a web service does that make this web service, RPA? When the RPA tool calls OCR to automate paper (or screen), is that no longer OCR? When RPA can be programmed to call amazons AI/NLP engine does that make that RPA, "Intelligent"? That's what some are trying to have you believe.

I also absolutely refute the idea that any RPA technology is a platform for Digital Transformation nor is it anywhere close to being an "Intelligent" Automation. I absolutely support RPA as an often important piece of these technology stacks and often being a bridge to getting through the last mile of "integration" even. This is why Pega brought OpenSpan, SAP brought Contextor and why more recently (I guess), Appian brought Jidoka. All of these acquired UI automation technologies (known as RPA) have/can more powerful when backed by vendors that understand what the next generation of application development has to look like; and how RPA can help.

2020 will flush this out I believe. Lets face it, the technology behind most RPA is as dumb as rocks (after all, most still have to send keystrokes and mouse moves to applications they have no real control over) and stop the user working at the same time. Some even still OCR the screen to read what's there (the oldest form of scraping I was doing back in the mid 90's). In fact, the technology is so old and freely available, this is why there are now 60+ RPA vendors who've popped up in the last 5 years who just use these technologies to create an RPA business/market/eco system. Can't blame them at all.

Ultimately, RPA can deliver you some great success and when done right, can be a powerful tool in your "Intelligent Automation" tool chest. I love it and with our unique RPA (Deep Robotics), have delivered many of the worlds largest projects that has delivered hundreds of millions of automated hours back to the business with it. Though many RPA projects across this industry still struggle to get any kind of scale and many have also shown up as well documented failures.

Regardless of success or failure though, the question is, "what next"? Now you've wrapped (band-aided) all your old stuff (tasks and/or processes) and made them more efficient, you are still left looking over your shoulder and saying, where is the next big innovation / transformation / optimization going to come from? RPA use cases dry up fast once you've got through the first few years and RPA vendors are looking desperate so beefing themselves up by hijacking any other piece of terminology. The "next" evolution of serious enterprise grade automation can ONLY come from Intelligent Automation, and that is NOT stand-alone RPA nor is it RPA as a back-bone for building next generation and must-have digital transformation.