Calculating the hidden cost of robotic process automation (RPA)

Danny Blog Picture-page-001

Calculating the hidden cost of robotic process automation (RPA)

Should you buy a kit car or the one in the showroom?

For companies setting out on their automation journey with a ‘build your own’ mindset, it’s worth pausing to examine the additional costs in detail, especially given the fact that the burden of mounting, related costs is the principal barrier to scaling automation strategies and delivering ROI. A report by HFS Research supports this view, it says that the software costs of RPA is only a minor portion of the total cost of ownership — representing between 25% to 30% of expenditure. Organizations are evidentially underestimating the hidden costs of deployments.

Let’s start by purely focusing on the running costs of a platform which underpins an automation capability.

In a world where the benefits of software as a service (SaaS) models for business applications are commonplace, the RPA market has remained surprisingly old school in its approach. Many industry-referenced deployments are hosted on premise, and as a result require experienced IT teams within companies to stand up and manage the necessary infrastructure which typically includes servers, databases and infrastructure licenses.   Even if the company’s IT teams work with third party cloud hosting firms such as Azure, AWS, GCE or IBM the underlying cost and complexity of managing the dedicated infrastructure that is housed in their cloud is significant.

Successful RPA deployments that are delivered at scale underpinning a wide range of core business applications will quickly be recognised as a business critical service. Add to that the productivity impact of removing even a single virtual worker who is delivering up to 15 times more workload than their human counterpart, and you can see the importance of paying due consideration to your RPA’s backup, resiliency and recovery practices.

Although reference architectures exist, these deployments do not unfold at the push of a button. Experienced consultants require time to craft a suitable configuration to meet your business needs and in the rush to move from proof of concept (POC) to Production deployment this can be overlooked and leave companies exposed to significant risk. 

As companies move to expand beyond the initial automation use cases they identified, they uncover how the behavioural traits of the RPA bots they chose can impede their progress.  While greater familiarity with these bots may highlight a few processes in which they can over achieve, limited cognitive skills will mean many processes prove too difficult.

Turning bots into intelligent digital workers is a challenge. The seemingly easy fix of engaging an ecosystem for bolt-ons and plugins to allow further use cases to be tackled, ushers in a new wave of costs. This panoply of technologies, all need their own infrastructure components, reference architecture, support terms and commercial agreements with new vendors.

Whilst a DIY deployment of RPA may suit a few organisations, the majority will find the cost and complexity of this approach puts a brake on their ability to derive business value from automation swiftly.  The RPA industry has matured and in weighing up their options before embarking on the next step in their automation strategy, companies now have the opportunity to adopt a SaaS model. In line with other business applications, their automation programme can benefit from the repeatable, scalable, secure, tried and tested SaaS deployment mechanisms, and grant internal teams the freedom to focus their attention solely on rolling out automation of business tasks and processes.

Anyone under pressure to reduce time to value of their RPA investment, will welcome the fact that in the timeframe it takes to agree network schematics, virtual workers can be already be being trained to automate the processes relevant to their business. 

Those that take a strategic, long term view will realise greater business results by taking an automation platform approach that is ready to drive, not ready to build.