Robotic Process Automation – will it be different this time?
The pros and cons of cognitive Robot Process Automation (RPA) have been brought to the fore in Jason Bloomberg’s article for Forbes, ‘Why you should think twice about robotic process automation’.
Here, Virtual Clarity founder and chief technology officer Rens Troost responds on why he believes you should learn from the past and maybe not invest your time and treasure in RPA.
We’ve been here before…
Back in the 70s and early 80s, when Graphical User Interfaces (GUI) emerged from Xerox PARC, IBM and some others invented something called HLLAPI (which resembled a beat-up jalopy, not just in name) by tacking on a user interface. The UI was just scraping that screen and trying to maintain tracking with a mainframe system that was moving target, and so it kept breaking.
Skip to Wall Street in the 80s to 90s, and there was plenty of screen scraping going on – a lot of market data that was only available on closed systems was made accessible as logical records via scraping, sometimes even to get around data licensing terms. Although there was a brief period when this behaviour was a real competitive differentiator and a good business outcome, it turned out to be completely unsustainable, and between 1990 and 1994, people invested in getting away from this screen scraping. Why? Because it kept breaking. Lots of early e-commerce companies also took this approach, as ‘Web Scraping’ took off and eventually crashed back to earth.
Between the late 90s and early noughties, there was a technology called Cicero that once again tried to save the world through screen scraping, promising more of the same – this time it was about orchestrating old applications to create new customer service flows... Guess what? It didn’t get far.
Now, RPA seems to be the same thing again – the latest in a long line of screen scraping approaches with a 40 year track record of “Meh”, except this time with even more yield-seeking investor dollars chasing the dream. What will be different this time?
Jason’s Bloomberg article lays out the game plan; sure, the robots are scoreless at half-time, but don’t worry, when they come out of the locker room, they will be artificially intelligent ‘cognitive RPA’ robots! Really!
The Mechanical Turk
In 2017, the business expense management app Expensify admitted to using low-paid humans to transcribe at least some of the receipts it claimed to process using “smart scan technology”. I don’t think this is unusual – I have witnessed companies whose “AI” is a bunch of offshore programmers writing not very intelligent stuff in response to rules triggers.
We’ve heard about RPA, it’s the darling of investors who wish to fire staff and automate “low value” jobs. However, according to a report by Enterprise Management Associates (EMA) to analyse RPA usage, users total 44% business and IT professionals at midsize and large firms, but they are not all enthusiastic about it.
According to Silicon Valley, these professionals are doing it wrong – but don’t worry, soon the smart robots will arrive to do it right. But why are they spending their limited resources on window dressing around processes they shouldn’t even be doing, instead of transforming their core so that it operates correctly?
RPA doesn’t change the game – it can make it worse
While RPA is sold as a small optimiser, freeing up resources to do the stuff you’d rather be doing, the reality is that most established companies are feeling the heat from new market entrants who are moving faster to deliver robust, appealing digital services. Our clients are faced with the knowledge they must deliver extremely pleasant digital services that people want to digest over their smartphones, and when they try to do that, if the core is brittle, the services keep falling over. Even if they are well-built by new, awesome people, the services don’t exist in a vacuum – they must connect with and be regulated in the same way as everything else the company has always done, as well as leverage the account information the company already has. We see this over and over again.
Enter RPA – it’s brittle in several ways. It scrapes screens, and, as we’ve seen, in a screen scraping environment, things fall over – we’ve known that since the 70s. Call me old fashioned! The way that RPA is being done today could only be invested in and purchased by a bunch of people with limited perspective. In ten years, I’ll probably be able to set up a screen scraping company and charm investors into parting with some their hard earned cash!
So, what should you do? RPA doesn’t solve the problem it thinks it’s going to solve; it creates and perpetuates the same problems you already have. Even worse, it lulls companies into a false sense of security and makes them feel that they are doing something strategic and valuable! And so, for most of our clients, I would advise them not to do it. There is far too much credibility staked, enterprise money spent, and investor dollars committed chasing magical AI thinking, when, in fact, it would be better to focus on properly transforming enterprise IT. Fix your core so that you can support your current and future digital environment.
Click here for more information on how Virtual Clarity can support your business through digital Transformation.