Robust process automation: thoughts on RPA

Last year the BBC published an “automation risk calculator” based on research carried out by Oxford University and Deloitte. It indicates the likelihood of computers replacing the need for certain human interventions and predicts that about 35% of jobs are at risk of computerisation over the following 20 years.

A robot doing a job in place of a human. Robocop might spring to mind. A slightly less sci-fi example can be found in your local supermarket in the form of self-service checkouts. It’s not that the robots are coming; they’ve already arrived.

Robotic Process Automation (RPA) is the practice of identifying mundane and repetitive tasks that are currently undertaken by humans, but which could be better performed by a machine. When I say “better” here, I mean more quickly and more accurately.

Consider the humble calculator. It has made simple mathematical processes extremely simple for the average Joe to perform. Suddenly you can multiply several long or decimal numbers together without having to get out a pen and paper.

73,324 x 9.29 = ?

Whomever the best mathematician in your office is, it’s unlikely that they will be any match for even the most basic calculator. Not only will the calculator be faster, it will also be more accurate. Suppose that you owned a business where it is necessary to do hundreds, maybe even thousands or millions of calculations a day. Even for relatively simple calculations, most supermarkets and retail stores don’t rely on their cashiers’ mental math skills these days; the cashier’s job is usually just to scan items, take money and input the amount into the till; the till does the calculation.

This begs the question: in the burgeoning age of RPA, why are some businesses still relying on manual/paper invoice processing when it goes against this trend of removing wasteful tasks?

I recently attended a major shared services conference where one of the big themes was how RPA has the potential to revolutionize the way shared services and outsourcing executives conduct their business. We spoke to both existing and prospective members of the Tungsten Network, and many of them told us that they view e-invoicing as a necessary (almost a given) step to best procure-to-pay practice. They know that they need to eliminate the manual touches associated with paper invoicing, but they also need more. The goal is to eradicate as many inefficient steps as possible, then to find smarter ways to handle the remaining inefficiencies.

Tungsten Network has the innate ability to 1) remove many of these inefficient touch points from the invoice approval process and 2) magnify the benefit of robotics by providing the timely, accurate data that is generated inherently from our e-invoicing network. The fact that we have compliant delivery that is proven across the world doesn’t hurt either.

Furthermore, while large multinationals often don’t have the luxury of clean data, Tungsten Network can deliver a globally scalable data overhaul that will change the way that robotics is applied. Tungsten Network normalizes data across enterprise resource planning systems, eliminates the occurrence of many invoice match exceptions and, with our development of Tungsten Network Analytics, has the capability to provide a holistic view of an organization’s spend that would be otherwise unavailable without significant data manipulation. This next level visibility allows businesses to identify trends and eliminate costs before they happen.

Robotics is already a fact of life and will continue to reinvent business procedures and practices over the coming years. Amazon shocked the world in late 2013 when it announced plans to begin testing a prototype of drone delivery. Many were sceptical, yet it’s now a couple of years later and Amazon are still going full speed ahead, while Google’s ‘Project Wing’ has announced that it aims to deliver packages by drone in 2017.

But robots are at their most valuable when integrated into processes that are already robust. Amazon became the model of e-commerce efficiency before drone delivery could enter the fray. If your organization is hampered by bad processes, the robotics answer will simply result in too many robots handling the same inefficiencies — potentially even creating new ones. Why not eliminate as many “RPA-eligible” processes as possible and then deploy robotics on the rest? The challenge businesses face is not whether to implement RPA, but when and how to best implement it given the unique landscape of their particular business. The ability to efficiently execute monotonous and repetitive tasks to a consistently high standard is too valuable. It cannot be overlooked.

Consider longstanding Tungsten Network e-invoicing customers and the large volumes of invoices they process on an annual basis. Conservatively, 60% of their PO-backed invoices auto-post for payment if they pass our pre-delivery PO validations. No need for exception handling or manual routing. No need for AP touches as the invoices move through the workflow. No need for RPA.

From there, our customers have only the remaining 40% to worry about removing inefficiencies through RPA. This narrowed scope gives them sharper focus, which in turn means they can precisely define what it is they need. And that means that robotic applications can add even more value, because they are being utilized efficiently.

Will RPA revolutionize your business? Probably, but have you looked into AP automation/workflow solutions? E-procurement? E-invoicing? Have you maximized your straight-through-processing thresholds? If so, robots may be your ticket to ultimate efficiency and savings. If not, we have some people you should talk to…