RPA series: Appian - creating an integrated robotic workforce with RPA

Srinivasan writes as follows…

RPA has become an invaluable tool for organisational digital transformation.

It has evolved rapidly over the past ten years and is now capable of automating human-machine interactions in the enterprise. For IT operations, RPA can effectively automate repetitive tasks such as updating system check-ups or running scripts on a regular basis. And now, that capability is starting to extend much further, both externally and to new services that IT needs to accommodate.

What is important is identifying and understanding where the high-volume, low-value manual work is and where human input can be removed; initially by automating the task using RPA, and then ideally modifying or removing the process altogether.

Across industries, including those which are highly regulated such as banking, insurance and pharmaceuticals, RPA is proving its effectiveness in bridging the gap between legacy and modern business apps.

RPA fills API EAI chasm

Where legacy platforms with limited to no API connectivity are prevalent, RPA is an increasingly important element of Enterprise Application Integration (EAI). If you were to take a medical record checking website as an example, where you can enter practitioner details to find information such as medical licenses, there is no easy system-to-system API that can execute successful integration. This would be an ideal use case for robots to mimic human behaviour and extract information without the need for the layer above (the user) to move out of the context.

Companies may find that their RPA engine has all the features it needs to build their process.

However, they may well encounter scenarios requiring logic that their RPA vendor does not have, which raises some challenges. Generally, RPA engines should have all the facilities to define business processes and business logic with the necessary structures in place to handle a wide variety of use cases – such as activities, events, gateways and integration services.

If this covers the majority of use cases, then the rest can be achieved by providing a functionality to expand the RPA engine’s capability in a scalable, maintainable fashion with a level of guarantee to ensure upgrades do not lose the functions being extended.

Doing so in a low-code fashion may increase the value of adoption.

RPA evolution path

RPA will mature in the coming years and most software will end up incorporating a degree of built-in automation or provide an RPA add-on to handle multiple scenarios. In addition, vendors will find ways to identify patterns and recommend methods to automate them. This is where machine learning and RPA will work closely together.

Work is already underway for RPA vendors to integrate with machine learning concepts that allow the bots to learn from their activities.

This currently falls under two main headings:

• Identifying where a process can be modified
• Recommending and/or performing the modification (based on the level of confidence)

The next stage in the evolution of RPA is to learn from already successful processes across the globe and accelerate innovation through collaboration. Particularly in automating the approval step, where RPA should recommend what are the likely steps to be considered, the exception scenarios, the handling of those exceptions, and how to involve humans in the loop.

This will accelerate utilisation of bots beyond the skillset and the individual developer.

RPA has already established itself as a vital tool of digital transformation for CIOs and CTOs with proven use cases. In the coming years there RPA will see further technological improvements, the automation of more manual processes, and ultimately, a move towards a genuinely digital workforce harmoniously integrated with both humans and artificial intelligence.