Australia’s Intelligent Automation: Failure to Launch?
Australia, like most developed economies, was quick to trial automation in its earliest incarnation as Robotic Process Automation (RPA). As some of the earliest adopters and poster children of RPA Australians enjoyed a sugar rush of quick benefits. Quickly though, many started to struggle to realise scale in their automation endeavours and many more stalled entirely with only a small number of bots automating simple activities.
A report by KPMG International and HFS Research, Easing the pressure points: The state of intelligent automation, shows less than 20 percent of Australian organisations claim to have an ‘at scale’ Intelligent Automation capability. Others, who have pushed on have found their bot armies difficult to manage without significant additional investment.
Despite the current reality, the potential for automation is huge. Automation was forecast to transform our ways of work with over 50 percent of our current workload being performed by machines faster and more accurately. So what went wrong?
It’s certainly not the availability of data. Data is increasingly accessible and usable, able to help systems make judgement based decisions. But still most organisations are barely off the launch pad, with fragmented automation capabilities struggling to deliver results.
The more advanced, Intelligent Automation needs an even more fundamental change of mindset than RPA. A range of organisational issues: process complexity, lack of business sponsorship and lack of IT architectural thinking have mired automation initiatives, greatly limiting their impact.
Based on discussions with Australian organisations, there are six themes that need to be addressed to succeed at scale.
Business – not IT led
This has been a theme since the early days of RPA but most companies did not really confront it as RPA was seen as a software rollout rather than a business initiative. In fact, 48 percent of organisations describe their Intelligent Automation initiatives as IT led. As a result, Intelligent Automation stalled with business leaders not fully engaged with thinking through how their operations should be optimised for automation. This includes better definition of processes and data management for machine based decision making.
End-to-end (E2e) approach
RPA gave many organisations an ‘out’ when it comes to this approach. RPA’s ease of deployment made it easy to automate several isolated steps of a process. Intelligent Automation requires end-to-end thinking as data needs to be captured cleanly and passed on through the process. Also, Intelligent Automation can automate even higher levels of a process that requires deeper thinking about whether this also requires organisational change.
Capture and control of intellectual capital
As automation increases, business has struggled with effectively capturing its IP. As more work is performed by machines, the relatively easy approach of asking a co-worker how a process works is lost to the organisation. We need to augment Intelligent Automation initiatives with ‘Intellectual Capital Control Rooms’, a capability to capture organisational business rules, be able to easily refer to them and modify them when required.
Many organisations thought of automation as RPA initiatives and picked one platform to drive their automation. This limits the ability for Intelligent Automation, for example dealing with voice or unstructured data, making judgement based decisions, or automating activities like digital marketing.
Only 10 percent of organisations consider their Intelligent Automation capabilities are integrated with analytics that feed automation execution. Intelligent Automation needs to be thought of as an architecture around core systems covering front, middle and back-office requirements.
Intelligent Automation technologies are evolving rapidly. Our experience is that every 18 months new technologies appear that can supersede prior technologies. Architectures need to be in place that allow for the easy substitution of new automation technologies into the automation backbone.
Cloud (decision) enabled
As intelligence levels of automation increase, the demands of computing power increase as well. This includes introducing new forms of data from additional third party sources or from IoT sensors, not simply scanned documents. The Intelligent Automation architecture should ensure it is cloud enabled – not just for the sake of providing cheap infrastructure for bots to run on, but to ensure that increasing levels of data are taken into account for automation decisions.
Intelligent Automation is becoming more common place with business executives more comfortable with analytics and automation as part of their roles. Australian organisations are looking for ways to align their automation aspirations with success stories from the few other organisations that have made it work. The reality is that this is not an agenda that can be driven by IT or business alone but requires integrated action across the organisation.
The good news is there are now enough success stories for organisations to learn from to reduce the risk of their own Intelligent Automation agendas.