RPA – better for individuals over entire business functions?
Fresh out of 2020 – the year of the pandemic and everything ‘unprecedented’, including the pace of digital transformation – and implementing automation in all functions of the business remains a priority.
When we talk of automation in the white-collar world of work, more often than not, we’re talking about the use of Robotic Process Automation Software (RPA).
Referring to software that automates repetitive, logical, and traditionally manual tasks, RPA acts as a bridge across systems, allowing companies to access several applications at once. It serves as a literal replacement for human workers, emulating the same keystrokes and mouse clicks as a human employee.
This technology saves time and money, boosts efficiency, and cuts out errors caused by human fatigue or distraction, and those benefits have led to the emergence of a vast and evolving RPA marketplace, predicted to hit a worth of close to US$30 billion by 2027.
But RPA’s rocketing success is arguably partly down to a misunderstanding over how the software should be deployed, and just how permanent a solution it can offer.
A band-aid solution?
“Early success with RPA implementation can foster a misguided tendency to apply it everywhere,” Robert Castles, chief technology officer at PMG, told TechHQ.
Of course, any technology whose purpose is to automate the processes established by former teams of human workers is not one that promotes actual digital transformation. In simple terms, RPA at its core is not ‘smart’, it essentially offers a script to automate a prescribed process.
Within this remit, RPA can reliably automate complex legacy processes. But if an application is updated or replaced, scripts will no longer function properly, leading to process failure or incorrect outputs.
Castles argues that RPA can entrench bad habits if used in the long term by stringing along with outdated technology and compounding existing silos between teams and systems. This is a “band-aid approach”, and it means that digital transformation gets pushed further away, not closer.
Indeed, according to research by HFS, 30% to 50% of RPA projects fail entirely. For those organizations that keep projects running, maintenance, and control issues are burning a hole in their pockets. According to Forrester, close to half (45%) of organizations suffer from poor customer support and engagement due to bot breakage on a weekly basis, or more regularly.
“The reality is that at some point, you’ve got to fix the issue,” said Castles. “And if a band-aid is there long enough, you’ll forget that it was supposed to be only a temporary band-aid. You’ll not only further entrench your technical debt for that legacy system, but you’ll add to it by having two systems to maintain instead of just one.”
In the RPA and automation space, the race is now on to make software that’s more ‘intelligent’ – that can learn and adapt, and change its processes in a way that optimizes the output.
Where RPA mimics human actions, Intelligent Automation (IA or IPA) aims to mimic human intelligence. Using technology such as computer vision, NLP, and machine learning, IA can automate processes that don’t have a rules-based structure, handle unstructured data, and provide answers based on subjective probability.
This is surely the direction RPA at large will shift towards, but it’s not to say that the software in its more pure form cannot benefit the business at an individual level of automation.
RPA for the individual
Thanks to the rise of low-code RPA solutions and vast improvements to user interfaces, RPA has become a tool that business users without coding skills, or ‘citizen developers’, can pick up and use quickly within the confines of the processes that they are most familiar with.
This is not RPA implemented at the company level, guided by a center of excellence, it is the use of automation as required by an individual. It can make lives easier and more productive, but by helping to sew automation into every little seam at the personal and departmental level, it can contribute to increased productivity across the entire business, while management and maintenance is handled by the user.
While businesses’ developers work on the implementation of big, organization-wide intelligent automation or upgrades to legacy systems, RPA can continue to be put to use day-to-day by citizen developers. Where uploading data from a spreadsheet to ERP systems might take 40 minutes manually, low-code RPA tools could enable a user to program the process and execute it in the background of more stimulating tasks.
Castles lent us another example: “Take the example of an insurance company with a legacy quoting system that runs on a mainframe. To establish a competitive market system, you want to create a web app with a good UX that customers can use to obtain quotes for purchasing insurance plans quickly and directly. How can you truly automate that process if you still need a person to ‘attend’ to the bots that get quotes out of the legacy system?
“To truly automate the end-to-end process, you’ll need a low-code orchestration platform that ties together all of the steps and systems – electronic signature, secure email, document management, etc. If you can automate the entire process and deliver results to customers quickly, you will stand out from your competition. If you delay, your competitors will get the market share before you do.”
All these micro use cases, day by day, can contribute to thousands of hours saved for businesses and a competitive edge as a result.
As with all technology, there are drawbacks to using RPA in this way at an individual level. Use cases and potential might be limited by internal teams’ reluctance to open up their systems’ APIs for integration, said Castles. Often this down to security concerns or policies.
Not only that, but data is frequently incompatible across different systems and applications.
Ultimately, RPA is indeed a quick fix for legacy IT and infrastructure at its core: “Fundamentally, it is short-sighted to think of RPA as the goal. It’s a means to an end for digitalization,’ said Castles.
“Since the lines between RPA and low-code are blurring with vendors on both sides incorporating the functionality of the other, the distinction between the two will become less relevant over time. It will then be up to the implementation teams, more than ever, to leverage the features of their selected technology in an optimal way.”