Building confidence in procurement
Robotic process automation (RPA) and artificial intelligence (AI) are two technologies poised to play greater roles in corporate procurement. Not only will they help transform the buying process into a more streamlined function, their use in payments will drive more effective control and compliance, and turn procurement into a department that adds strategic value to the business overall.
That’s the view of Andrew Burgess, author of The Executive Guide to Artificial Intelligence. “From a maturity point of view, RPA is a hot topic and it’s rare to find a large company that isn’t looking at it in some way,” he says.
The human impact of RPA
Realistically, it’s RPA that will have a more immediate impact on payments, not least because it addresses the need for efficiencies in payments and other procurement processes. “Wherever there are people carrying out rules-based computer processes, there’s an opportunity to put RPA in place,” Andrew says.
“The software itself mimics what people can do on computer systems, and in processing that usually means crossing several systems,” says Andrew. RPA replicates the process to provide benefits of cost and capacity. “It’s cheaper to purchase an RPA licence than to employ a person, and RPA can work 24 hours a day, seven days a week, 365 days a year, whereas a human can’t,” he notes.
Robots also tend to be faster and, assuming the software has been configured properly, they are more accurate. In effect, RPA assures compliance in payments and helps procurement eliminate uncontrollable or rogue spend. “RPA gives you the comfort that everything is going to be error-free and auditable – you can always look back and see how a process or decision happened,” Andrew adds.
“RPA gives you the comfort that everything is going to be error-free.”
Mundane tasks that do not add any real value to procurement can also be eliminated through RPA. In payments, those tasks could be in exceptions, some routing processes, reconciliations or supplier onboarding. Much of the latter, for example, is simply entering data or sourcing data into or from multiple systems. “RPA could do this instead, enabling procurement to build efficiencies,” says Andrew.
Not only could businesses realise cost savings as a result, but employees’ time could also be freed up to do the more value-added tasks such as building supplier relationships.
Managing risk with AI
While RPA can provide efficiency benefits to payment and procurement, it is AI’s self-learning capabilities that could bring additional value to both in the future. “Not only can it create efficiencies, it also gives you insights and additional decision-making capabilities you wouldn’t normally have,” says Andrew.
One area AI could help in the future is risk assessment. A traditional approach may use four or five different criteria to determine whether a supplier is a potential risk. Instead, through machine learning, AI could work through a much greater range of data to decide whether that information should impact a decision positively or negatively. “As long as it has enough data to be trained on, it can help identify those suppliers that might be more of a risk than others,” Andrew says.
“AI gives you insights and additional decision-making capabilities you wouldn’t normally have.”
In this case, a static AI system could be set up, with reviews scheduled in at pre-specified intervals. Fraud, on the other hand, is an area where a dynamic AI system could be more suitable.
Andrew explains “People frequently come up with new ways to defraud and the sooner you spot fraudulent transactions, the sooner you can stop them. When using AI, you would want to make sure the technology is learning on the job all the time. Using a dynamic AI system in this case is more complicated and requires more processing power, but also means it could learn very quickly as things change dynamically.”
Businesses and procurement departments adopting both RPA and AI technologies need to be cognisant of the different roles they have and will have. Andrew says “Where you need compliance and zero errors, and you want the same thing to happen every time, that’s where RPA comes into its own. Whereas if you’re looking at confidence levels, then AI is much better suited. You should get the right technology for the right situation.”