In this guest post, Procurement Leaders invites Tungsten Networks' Stefan Foryszewski to give some insight into where technology will change the quality of decision-making in the procurement function.
Getting bang for your buck is one of the biggest priorities in procurement today, especially when supplier numbers grow ever larger. For many years, spend analytics software has been helping procurement departments to identify where cost savings can be made. But now we're at a turning point where the technology is coming of age and reaching new levels of sophistication.
With the rise of e-Invoicing and e-Procurement, the amount and accuracy of data available to procurement teams is growing. When staff are based across multiple locations and in different departments it's easy for duplications to occur. This means that teams can miss crucial opportunities to negotiate on consistent pricing and economies of scale, which could add up to hundreds of thousands of pounds. To give one example, recent analysis of UK's NHS spending identified that while some trusts paid less than £4 for a box of needles, others paid £31.68. Clearly, if this was identified sooner, huge sums could be saved to the public purse.
We see spend analytics as a crucial area to reduce spend and improve efficiencies in procurement. As a result, we have recently partnered with Goldsmiths University in London to launch the Tungsten Centre for Intelligent Data Analytics. A dedicated team of academics will be charged with researching and developing our spend analysis technology, Tungsten Analytics, to enter the next realm and crucially, use state of the art artificial intelligence to do it.
The term artificial intelligence conjures up associations with science fiction films, but in the next generation of spend analytics technology, it will help to solve a number of very contemporary problems.
Firstly, helping computers to understand and interpret written text. Machines can't read so they need to find a way to identify semantics. For example, if two product descriptions are written in slightly different ways, – say ‘50ml syringe' versus ‘syringe 50ml' – how do they identify them?
Next, spend analytics will be further developed to learn functional equivalence. To use a car analogy, a human would know that a Ford and a Nissan are both brands of vehicle, but a computer wouldn't necessarily. The next generation of spend analytics will see computers programmed to learn these subtle differences, so that better comparisons can be made.
Finally, we can expect the technology to evolve to achieve greater levels of trend analysis, using financial modelling to predict future pricing patterns and to assess supply chain risks.
Benefits for business
Intelligent computing and data analysis have uses across the business world and major global firms are sitting up and taking notice. Google and Facebook are investing heavily in research and development in these areas, while Amazon Web Services has set up a dedicated Machine Learning team. This is a growing industry that is being realised by business as a highly lucrative area.
But spend analysis is not just for big business. While large firms have more data to analyse, there is often more potential with small and medium sized businesses, which don't already have efficient purchasing processes in place. When businesses grow rapidly, they can have a tendency to pull in suppliers from multiple directions and it can often take years for consolidation to take place.
Equally, for big businesses, even small cost savings add up to a big figure and significant benefits can be reaped from closely analysing spend. Identifying these cost savings is one of the biggest challenges facing procurement teams today, but with technology developing quickly to assist with these decisions, help is at hand.
Stefan Foryszewski is executive vice president at Tungsten Network.
This contributed article has been written by a guest writer at the invitation of Procurement Leaders. Procurement Leaders received no payment directly connected with the publishing of this content.