There aren’t too many people who can boast to be one of the true pioneers of the emerging robotic process industry, or whatever we end up calling it, but one man who didn’t need to recreate his resume in 2017 is Francis Carden, one of the original brains behind recreating desktop automation to become Robotic Desktop Automation (RDA) which is such a key component of the broader RPA offerings attracting so much noise and attention today. And he now leads the whole robotics strategy for customer engagement and data orchestration giant Pega.
Francis – a British Floridian these days – is never one to hold back, and when one cynic mercilessly described some of today’s RPA rollouts as “lipstick on a pig” at our recent New York FORA event, he just had to take it one stage further to declare them as “lipstick on a pig’s arse”. So without further ado, let’s hear from the granddaddy of RDA himself…
Phil Fersht (CEO, HFS Research): Francis – firstly, tell us about the OpenSpan business you co-founded and how this evolved to the acquisition by Pega?
Francis Carden (VP, Digital Automation and Robotics at Pegasystems): In 2004, a group of “tech head” advanced windows operating system engineers approached me about forming a company using an automation technology vastly different from anything proceeding it. It let companies rapidly automate, through the Graphical User Interfaces (GUI) of windows applications, any task or process that a normally would be done by a human on their desktop.
On first hearing of it, I had to laugh and said to myself, “It sounds like screen scraping.” I didn’t believe what they were proposing was even possible. But I did a deeper investigation into the background of these engineers and the technology and found it both unique and incredible. So I bought in – and even invested a significant amount myself. Today, we label our RPA technology “Deep Robotics” given it is so different.
In 2005 we started OpenSpan and over the next three years, as CEO, we convinced a consortium of Atlanta technology angels and then attracted four Tier 1 VC firms that what we had was something different. The customer base grew rapidly. Our patented technology started life as being called “Surface Integration” because it was so different from screen scrapers of the past (and current RPA). But we quickly settled on “Desktop Automation” (known today as RDA), and “Automation Broker” (known singularly today as RPA). We ended up implementing most of the world’s largest RPA projects over the next 10 years.
Then Pega acquired OpenSpan in 2016. Pega recognized our RPA approach was markedly different from any other RPA vendor. After we got talking, I realized our unique technology would be even stronger if it were part of something “bigger.”
What I mean by that is organizations on the digital transformation journey shouldn’t aspire to RPA as the end game. You really need to look at RPA as a short-term plug to fill the gaps on the road to get there. When you deploy RPA, you’re just masking the poor processes behind it. And when application UI changes, the bots are prone to break. It might be weird to hear me – an RPA vendor – say that. But it’s the truth.
And that’s where Pega’s digital process automation (DPA) software comes into play. Pega enables you to digitally architect those processes the right way, from the ground up. No more organizational silos, no more inefficiencies, no more needlessly manual processes … and really, no need for RPA to always be the first option. RPA gives you a digital automation jump start, and when it’s time to deploy DPA, those bots can be put out to pasture.
And the good news, at least for us, is that everyone is on a digital transformation (One office as HFS call it Phil) journey, and it really never ends. RPA and DPA are the perfect one-two punch. So the tie up and combined value prop with Pega was, and has been, an ideal match.
You’ve been the undeclared granddaddy of RDA, so what is the key “difference” between RDA and RPA… unattended and attended… etc?
Imagine if every compiled windows application ever built allowed itself to be automated from within its own code, robustly and in real-time. You wouldn’t need any RPA, right? However, very few apps running on windows are easily automatable at the UI layer – at least, without resorting to screen scraping techniques like MSAA, UI automation and OCR.
In 2005, we patented a more sophisticated approach to UI automations – a “plug-in” like architecture that enabled any compiled windows application and all its objects to be non-invasively automatable in a single architecture. This “deep robotics” automation occurs inside the windows layer, inside the memory of the machine, and runs as if the original developer of that application had natively built it. This plug-in like approach automates at the application’s original speed (10x to 100x faster than scraping), highly robustly, and far less susceptible to application UI changes. More importantly, and hence the dual name “attended RPA,” it even allows the desktop worker to use their machine, keyboard, and mouse while the automations are running. This is something unheard for old school scrapers and other RPA products. It is this pure form of RPA that also enables it to be used as RDA, giving every worker their own personal robot or “co-bot” to automate from one percent to 90 percent of their work. Agile RPA. And only then, after you’ve rolled out RDA, should you look at those same processes ready for full automation with RPA (unattended as opposed to attended) operations.
Before the Pega acquisition and continuing since, we have 100s of 1000s of bots deployed across some of the world’s largest enterprises in all industries. By the way Phil, being a Grand-daddy makes me feel old, but I did calculate that I’ve been doing automation of the UI for 20 million minutes of my life!
So what happens next as RPA (assume RPA includes RDA here) moves from “tinkering” to broader enterprise adoption – where do you see clients finding the most value, and how can they scale their people and tech platforms to accommodate?
Good question, Phil, and the key word to the question is “tinkering’ – which is how most RPA vendors excuse themselves, after 10 years of trying, for not getting many customers to really scale or still stuck in pilot stages.
It is hard to believe, looking at the crazy RPA vendor valuations today, but yes, most RPA projects are tiny compared to scope of the company’s using them. RPA is not new, so the real question enterprises need to ask is, “What’s stopping this large-scale adoption across the globe, and why do RPA vendors keep insisting it’s because it’s new”
Attended RPA (RDA) scales — that’s a fact. It automates all the easy stuff in an agile way. But RPA unattended constantly struggles with trying to automate everything, both the easy and the hard stuff. Putting RPA band-aids on top of old and tired processes is just wrong for the long term. As Gartner says, “RPA is a tax on legacy,” but businesses are often so enamored by the hype.
But the thing is, RPA will indeed scale if it’s part of something bigger. RPA is tactical to the extreme. Digital transformation is strategic. What happens though if you combine the two? This is where software robotics shines. Using RPA and RDA to plug gaps that currently prevent digital transformation allows business and IT to align to solve the real problems. This gets you real and rapid ROI from RPA out of the gate, but equally, it encourages planning to then “fire” the robots as fast as you deploy them. Not something you’ll hear other RPA vendors promote Phil. With robotic automation capabilities embedded and fully integrated into the heart of Pega’s DPA, we are seeing enterprises really changing the way they compete in a race to become them most digital company in their market(s). Analog companies, using tactical RPA only as the glue, or those not buying into your one-office analogy, will simply not survive.
Francis, we’re seeing some tentative moves from “big iron” ERP vendors such as SAP into the process automation space, but do they really want to delve into this world, or are they merely ticking the “we have an RPA module” box?
Like I said, if all new applications were built to be truly digital and open, then there is no need for RPA at all. But we’re living in the real world here. While we wait for that to happen, the real “transformation” vendors selling the digital dream need to be able to use tactical technologies like RPA to help their customers plug the (hopefully) short term gaps. This gives customers quick relief where they are strangled by their legacy systems with no other integration capability. If a vendor doesn’t have this kind of automation, this sets the customers up for failure – they need to automate now to compete, not six or 12 months from now when your big overhaul project is finally done.
There are now 30+ RPA vendors, but I think the bubble will burst on these companies riding the hype wave. Most of them use many of the same old scraping technologies, and none of them have a unified DPA play to help companies for the long term. I’m not sure how long these RPA companies will be able stand on their own. Consider that Pega now includes bots as a standard capability of our DPA platform. We’ll give you all the bundled bots you need as part of your DPA strategy. So yes, we’ll likely see more consolidation in this market.
And how do you see the “blending” if classical RPA with some of these newer AI products on the market? It feels to me that RPA has really evolved from the process/operations side of the business, while AI is some magical vision being pushed hard by IT, but lacks in real business applications? What is real versus fantasy in your view?
This is the biggest myth of RPA. The idea that AI will help accelerate getting RPA to scale. It’s a red-herring perpetuated only by those in the RPA industry. No bias there then!! I concur with you Phil that AI is central to becoming a truly digital company. However, the idea of using AI with RPAdoesn’t fly – at least not right now. AI must live central to anything an enterprise does, everywhere. AI needs data, and LOTS of it. But very little of that will come because of RPA.
In order for AI to be really valuable, it needs data from as many systems as possible as well as data feeds from interactions with as many customers as possible, in real-time or extremely fast. Only then can you use AI start to predict what customers are going to do and/or create models that make the best decision for each individual customer. RPA would likely touch less than 1 percent of 1 percent of that data, so if anything, the RPA world is embarrassing themselves by trying to join these two tactical vs. highly strategic dots.
Don’t get me wrong, AI can create work for RPA bots, and maybe some RPA feeds some data into AI, but that doesn’t make RPA Intelligent. At the heart of AI is a digital company with data feeds from every source feeding into it. The DPA (one-office) movement is joining real transformation technologies together, acting as one, with AI at the center. The vendors that have DPA are the ones delivering IT and business on the promises of the past – a real digital company – and envy of their peers!
Thanks for the insights, Francis… we’ll be watching Pega closely in this space this year as the market finds its sea-legs… and some more attractive lipstick!