Dude, where’s my personalised spend report?

COVID-19 has dragged many kicking and screaming into the digital era. The Queen has been seen on Skype, the UK Cabinet met on Zoom and many of us mere mortals have turned to online shops to get essential items. Mintel reported the online grocery market in UK is forecasted to grow by 33% in 2020, up from a growth rate of only 2.9% in 2019.

Nick Carroll, Associate Director of Retail Research at Mintel, said: “Over the course of just a few months, COVID-19 has had a seismic impact on Britain’s grocery sector. The pandemic is giving a significant short-term boost to online grocery services…, however, the impact will last beyond the crisis.”

Will COVID-19 have the same seismic impact on procurement?

Some organisations identified the benefits of digitisation before the pandemic. Vodafone, for example, embarked on the digital transformation journey a long time ago and are more efficient and have higher levels of compliance. Better visibility helps procurement professional identify opportunities and issues more quickly. A greater use of artificial intelligence, and in particular the use of one of its sub-areas, machine learning, improves productivity.

Despite improvements that brought about the recent “AI summer”, the Economist believes that progress might be reaching a plateau. The forecast is based on an analysis of the factors that have created the latest machine-learning revolution, viz: improved algorithms, more data, and more powerful computers.

Machine learning uses thousands of examples to train algorithms. The resulting systems can do some tasks, such as recognise changes in the pattern of third party spend, far more reliably. Systems programmed in the traditional way with hand-crafted rules often fail to fully incorporate key factors, like changes in the cost centre structure. Machines, however, are not “intelligent” in the way that most people understand it, a phenomenon known as Morovec’s paradox. Machines excel at well-bounded tasks but get things wrong if faced with an unexpected input, such as a pandemic.

Despite digital systems collecting every increasing amounts of data, some critical data fields are often absent, or require further data wrangling to get right. When supplier master data contains duplicates, for example, the machine is unable to provide an error-free spend report.

AI systems’ demand for computing power is expensive. Building a business case, competing with other areas of the business for the same resource and implementing systems takes time. Given the downturn in the economy due to COVID-19, many businesses will not have extra cash to invest.

If COVID-19 has a silver lining it will be to jolt procurement laggards into action and use the “AI winter” to catch up.

If you enjoyed this article, please consider sharing it!
Icon Icon Icon