Robotic Process Automation – Fad or the Future?

There seems to be a lot of confusion about Robotic Process Automation (RPA), yet it has been around for some time now. Some people think the workplace will be invaded by androids. Terminators will be pushing us aside to tap at our keyboards. The reality is somewhat different though. RPA is overwhelmingly deployed as software robots to process repetitive tasks at far higher speeds and with greater accuracy than humans.

The key here is that to realise any benefit, the task has to be repetitive to realise value in automating, and simple enough for the technology to perform reliably. What is considered simple is changing as time moves on with the technology becoming more capable and incorporating AI technologies.

Tasks like accounting, procurement, payroll are excellent candidates for RPA as they typically include tasks that are relatively easy to capture as a workflow, with high repetition rates which increases the value of the automation.

Who are the big players?
Although the industry is relatively new, there are a number of major players in the market including Blue Prism, UiPath and Automation Anywhere. SAP has also invested heavily in RPA, enabling built in automation for SAP systems. By incorporating RPA technology into the SAP applications and databases, the automation will become easier to implement and more tightly integrated with the applications.

Pros and Cons?
Simply put, RPA will enable businesses to reduce costs by freeing up employees from drudge work into tasks better suited to humans – tasks requiring creative thought, initiative and the ability to deal with variation.

By the end of 2020, Gartner is predicting RPA and AI technologies will free up 65% of employees in business shared services centres and the RPA market itself will grow to be worth over $1B.

There is a potential dark side to this bonanza though – jobs losses and stifling of application innovation are some of the issues. Magical success for a deployment is by no means guaranteed. Careful use case evaluation and selection is very much required. Project oversight, change management and governance is essential.

Who is using it?
I thought I’d talk to an actual RPA developer to get some insight to real world application. My son happens to be employed by an Australian government agency where he has worked on automating their systems.

What sort of systems have you automated with RPA tools?
The systems I have experience applying automations to are; SAP payroll and HR systems, SAP Child systems, and Outlook.

Where do you think RPA technologies help the most?
The best candidates for RPA software in government typically are HR systems and financial reports. The generation of invoice management reports and budget reports is one of the biggest time-consuming requests the agency handles on a regular basis. The implementation of automation in these areas is massively beneficial as the returned massive resources to the business unit and added improved perception from clients.

RPA in government is most successful when implemented on processes that are already defined and not undergoing other development.

Where do you think RPA technologies don’t help?
RPA implementation struggles in areas that are already undergoing change, such as departments effected by the recent Machinery of Government (MOG) change that has huge process and policy changes.

In my experience automations are quickly scrapped where the system is still undergoing development and in-system changes adversely affect surface based automations.

What is the hardest thing about developing RPA systems?
The biggest hurdle when developing in-system automation is process definition and managing stakeholder expectations.

If the process has other changes ongoing it is always best to define the process after the changes are implemented. Any developer knows the pains of creating something that can’t be included as part of an update due to compatibility issues. In short, develop completed systems – there is no point working with a beta as the entire platform may not work with the finished automation.

This can be hard to do in government as the push to roll out updates in-parallel means managing stakeholder expectations on delivery timeframes. I would advise all developers to stay in the loop with all system development meetings and engage with local subject matter experts at every available opportunity.

Are there any downsides to using RPA technologies?
The only real downside to RPA is the perception that staff have of the new technology potentially “doing them out of a Job”. This is never the case however, in every successful case of in-system automation staff are able to get new training and new experience to further their careers. Although the perception can make implementation of the change hard, once it’s done the perception is quickly turned around.

What was the hardest thing to get right?
One of the most important and most critical aspects of RPA is performance capability mapping – you need to know how well the process currently performs and how the automation is expected to perform.

As with every system it needs to be capable of handling increased workload with no effect on reliability. So, when undertaking the development plan for new automations, make sure to investigate the systems limitations and issues with current functionality.

It’s always best practice to improve on reliability when implementing automations – an intelligent automation with the ability to self-diagnose issues and recover from errors is more valuable than an automation that requires constant intervention.

Finally, where do think RPA technology is going?
RPA technologies are being implemented in almost every modern business or government around the world.

It is quite literally the way of the future.

I expect that by 2025 we’ll see RPA established in every government department and as the technology is implemented, we’ll see the perception totally change.

The Future
As my son mentioned, RPA is going to become pervasive in modern organisations. AI technologies are increasingly being included in RPA, enabling an easier, less proscriptive way of designing the workflows.

SAP is investing in this future with an intent to bake intelligence in to create an integrated automation platform.

Of course, with less human direction comes more risk of the wheels falling off, so even more time needs to be spent in the oversight of the project. And of course, the potential negatives mentioned above are amplified.

The end state of this is replacement of the human oriented applications altogether, allowing the machines to exchange data directly. Perhaps we’ll call that system Skynet 😊

Like it or not RPA is here to stay (until it is superseded or made irrelevant). I think when that happens I’ll go get a beer – I wonder if I can automate that process?

This blog originally featured on Linkedin.