Back to Hub

Clarifying Robotic Process Automation: What It Is (and Isn’t) for Procurement

01/08/2018 By

Image by joebakal sourced from Adobe Stock

Robotic process automation has become a buzzword most procurement professionals can’t escape. But despite its seeming ubiquity, RPA still tends to produce just as much head scratching as efficiency gains.

Taken out of context, RPA can seem like yet another disruptive technology to keep track of alongside machine learning, blockchain and the like. Really, RPA is a single element of an overall procurement digital transformation strategy, and one of the more readily accessible technologies, in fact. For many procurement organizations, adopting RPA can be a first step on the digital roadmap that helps sustain more complex initiatives further along.

To help you get started on that journey, here’s a quick guide to what RPA is — and what it is commonly mistaken for — as well as some examples for how it can be applied in both practical and advanced contexts.

Defining RPA

Robotic process automation is simply a type of software that mimics the activity of a human carrying out a task within a process. The term is derived from the idea of a “software robot,” an application that replicates the actions a user takes through the user interface of a computer system.

Said another way, RPA records anything a user does through clicking on a screen or typing into a keyboard and then repeats those actions as directed. Software robots can complete tasks such as opening emails, completing e-forms, and recording and rekeying data into a spreadsheet.

The benefits of such a technology are obvious. RPA can be used to execute repetitive, often clerical tasks more quickly and accurately than humans can. It is frequently used to input and transfer data between applications, such as retrieving invoice data from a PDF reader and entering that data in Excel.

What also makes RPA interesting is that it is accessible for most procurement professionals, because RPA applications require no coding to create. Using specialized software (e.g., Blue Prism, UiPath) users design the process flow and teach the computer to mimic the usual steps in a business process as seen through the computer’s user interface.

Common Misconceptions

The term robotic process automation sounds highly technical, so it’s not surprising that RPA is often confused with machine learning or advanced artificial intelligence. While ML and AI can be used for the purposes of automation, they take a different approach.

Machine learning uses statistical analysis to “learn” from reams of data and teach a software application, using these lessons to make highly accurate predictions about specific types of data. Artificial intelligence allows software to learn and interpret information, such as text through natural language processing, so that an application can communicate with and act in place of the user. This culminates in the emerging domain of cognitive technology, which attempts to replicate the human capacity for analysis and providing predictive insights by creating an independent, thinking machine.

RPA does not attempt to read, interpret or think. Instead, it simply replicates whatever actions a user records in the application. RPA is the form of automation that most any procurement user can implement quickly and easily, allowing an organization to get its feet wet in the digital world without taking on more than it can handle at first.  

Applications in a Procurement and Sourcing Context

The classic procurement use case for RPA is in transactional processes such as invoice receipt and validation.

When a vendor sends an invoice, the RPA application will open the email, open the PDF invoice, retrieve the relevant information, log in to the ERP and input the data, and possibly check that data against a PO (which should already be in the ERP) to make sure the invoice is accurate. If everything checks out, the application could even issue a payment.

While this is a simple example, the potential benefits are immediately clear. Retrieving and entering this kind of information, especially in a large procurement organization, is tedious and time-consuming. Not only could a practitioner’s time be better spent interacting with stakeholders or working on more strategic projects, but after a while, the user is bound to make a mistake out of fatigue or boredom. Implementing RPA for such processes quickly and easily addresses both issues.

But RPA’s usefulness is not limited merely to automating clerical business processes. In its most advanced form, RPA can change the way sourcing negotiation is done — for both procurement and suppliers.

In certain sourcing situations, software robots could conduct the supplier-buyer negotiation from both sides. The scope of goods or services requested and offered would need to be determined ahead of time by humans, but the process of requesting bid information and transferring this information back and forth until the qualifying bid is selected can mostly be left up to a robotic process.

This is not the proper strategy for a major category redesign, of course. But for lower-stakes sourcing exercises, such as consolidating tail spend with multiple vendors into one supplier, procurement could certainly benefit from reclaiming the time it would usually spend babysitting bidding and negotiation. That time could be spent instead on the big sourcing projects that catch the eye of the C-suite, while allowing procurement to also claim the quick wins that add up to obligatory savings goals down the line.