To RPA or not to RPA.........

Robotics Process Automation, better known by its three-letter acronym, RPA, is a promising solution to a few process and system issues than an aging system experiences in a real-life situation. The promise ranges from fixing the to leading the way into next-generation AI [Artificial Intelligence world].

In this discussion, I aim to address how organizations can leverage RPA, strategically as well as tactically to address their process automation needs

Robotics process Automation in its simplest form can replace a task, which is rule-based and repetitive in nature by a bot.In the age of sci-fi movies one is tempted to imagine physical robots of the size of a puppy at least if not humans, doing rounds on an operation floor of an organization, but with due apologies for disappointing the imagination, a bot in question is a software program. A bot can at best be compared to a MACRO functionality in an excel sheet. This “MACRO” however is a little more advanced and is capable of working across systems and applications based on the pre-configured rules. This functionality of being able to traverse across systems and applications allows a bot to replace humans for any repetitive, rule-based mundane task, bringing in machine efficiency and accuracy along. A bot can monitor a work queue and perform a task it is programmed to do based on the queue trigger, it can monitor a mailbox and play a mechanical role in the workflow chain.

Apart from obvious advantages of efficiency, 24/7 availability and accuracy, RPA offers few added advantages. It can be configured quickly and is ready to be deployed in 4 to 6 weeks that too at a fraction of the cost of a software patch with no regression impact on the applications involved. RPA is system agnostic and cost-effective replacement of human effort employed for rule-based repetitive tasks.

An RPA bot at this stage and for this discussion is not capable of doing tasks where decision making is required, it can help you calculate the fields of a defined scorecard but it is not capable of branching out an action based on an outcome, it can help you with fields to input into an optimization model but select an optimization outcome. RPA is neither cognitive neither employs neural networks. RPA is strictly rule-based, so if rules change you will have to reconfigure the bot.

Having scoped out RPA, let us move to discuss “what is it good for………….”!

RPA can surely patch up an existing system, which has, by virtue of aging of deficient initial design, have moved farther away from a process. The gap created by this shift over the years is currently managed manually with a possible UDA [User Defined Application read EXCEL sheets…]. In the world of regulatory control and constant compliance scrutiny, RPA seems to be quickest and cheapest remedy to fill these gaps. RPA can fill these gaps easily can be used as a perfect tactical solution too, however, process strategists should look beyond the obvious immediate term fixes.

The saves from reducing FTE [Full Time Employees] may be an immediate lure to adopt RPA and also proves to be a defendable reason to project a handsome ROI[Return On Investment] while getting the projects approved, this is just a very myopic view of the benefits. FTE saves at the best is a low-value byproduct for the organization which does more harm than good to the overall culture.

The fact that RPA can plug leakages and increase the operational life of an aging system should be seen in conjunction another fact that AI [Artificial Intelligence] is staring us quite in our faces. The statement that eventual AI based solution is the future is hardly a prophecy now, and the speed at which things are coming, we may already be late. AI is not a case of “IF AI” but that of “When AI”.

RPA can be used to plan and translate business requirements on a working system at a fractional cost over a period of 2 to 3 years. This working system with an aid of RPA serves as a live simulator for your future business processes. While RPA keeps the process running, process planners should tick the brans of business users for long-term expansion and translate [if possible] it into processes, experimenting on a live system using RPA. Needless to say here that all this needs to be documented and put design principle rigor.

The exercise may seem an advanced form of requirement gathering, which it is, but it is a lot more than just that. This exercise will allow business and process designers a live sandbox experience, this will also allow the planners to have a view of AI-based change they will have to deal with. This will bring in the AI change based on the organization pull rather than following an imagined end state. This will prepare the organization and its existing workforce for the imminent AI shift. In RPA can be used as a tool for preparing the ground for AI landing with lesser risk, cheaper cost, and greater certainty.

AI as a tool is a huge shift from any change we have seen in recent years, and like all changes, this will also come with its set of shortcomings. The AI is not going to remove human aspect completely in a flash. The organization to be ready to reap the maximum benefits by using AI as a strategic tool, the defining factor in this is going to be the turn around time and depth of penetration.

To conclude organizations should focus on a long-term strategy while presenting a business case for RPA. The benefits and ROI too should be viewed beyond FTE saves and process quick fixes, with a long-term view.

Ashish is a change and transformation leader, with a deep interest in emerging technologies and its usage in Financial services processes. The views expressed above are his personal based on his deep experience in managing transformation programs in various banks he has worked with.