In the last couple of years, the market for robotic process automation or RPA software has clearly been booming. Yet for all the apparent success of this promising new business automation technology, a certain sense of disappointment has also crept into the market already, as RPA faces a number of common issues and challenges. Time, then, for a quick look at the main reasons why RPA projects fail - and how to avoid them.
As with any emerging new technology, a lot of the disappointment is due to poor expectation management. According to Gartner, at the end of 2018, RPA was not only squarely on its way to mainstream adoption, but it also resided at the very ‘Peak of Inflated Expectations’. This has led some companies to fall into the trap of overpromising and under-delivering on their RPA projects.
It’s the strategy, stupid!
The lesson to be learned here, obviously, is never to assume that simply implementing a new technology, such as RPA, is all that is needed to achieve a great ROI. As always, it is not enough to buy into the new technology, there has to be some sound strategy behind it too – one that makes definite sense, if it is to be at all successful.
Furthermore, any new technology really has to fit in with all the different aspects of your organisation, from its IT infrastructure to its culture. Otherwise you run the all-too-real risk that it becomes a standalone business function which cannot meet the collective needs of your various users.
It’s the business, stupid!
There are other factors at play too, though, apart from these overblown, unrealistic expectations of the ROI that can be achieved through an RPA implementation. In fact, at the end of the day, RPA projects do not differ all that much from other IT projects, in that the main factor of failure is more often than not the lack of a clear business case. So, in order for your RPA project to succeed, you have to optimise it to your specific business needs.
The risk of failure is even greater still, if your RPA project is not business-led but IT-led. RPA solutions are not SaaS- or cloud-like services that work immediately, out of the box, with little to no customisation. Therefore they require the buy-in of key executives in order to get deployed and properly maintained for sustained operations.
It’s the process, stupid!
Robotic process automation, as the term itself suggests, allows organisations to automate certain business processes – mainly repetitive, manual tasks - through the use of (ro)bots. This improves their operational efficiency, potentially saving them substantial costs. It also gives their employees the opportunity to focus on other, non-repetitive tasks with (more) added value, leading to increased job satisfaction.
Of course, all of this presupposes that you target RPA at the right processes. Unfortunately, that is not always the case. The process requirements or target criteria for RPA to actually work and be effective are fairly straightforward, though. First, the process you wish to automate needs to consist of a series of stable and repetitive steps – a highly structured workflow, if you like, such as a fixed sequence of manual clicks. Secondly, that process needs to involve structured data, preferably high volumes of it. And last but not least, you need to be able to translate the process into a simple rules-based software. If these three conditions are met, in my experience your RPA project has a really good chance of succeeding.