Fools rush in where angels fear to tread

But we can successfully bring AI to procurement

Warren Buffett stirred this year’s Berkshire Hathaway shareholder meeting by saying AI scams could become ‘the growth industry of all time.’

More seriously, Buffett also acknowledged that AI has enormous potential for good as well as harm, saying it is much like any other technology humans develop.

So, while scammers are real, and there are concerns about AI running away from its creators, there are still very real, practical, even mundane, business applications for AI that make perfect sense – entirely safe sense – that we ought to be acting on, because we have a duty to operate our businesses more efficiently and effectively.

Moreover, there is also a competitive imperative to be thinking about how AI might be used more strategically.

Momentum is building fast

Since the arrival of ChatGPT, the world has moved fairly quickly on AI, yet its roots are deep and go back to the 1950s. For generations now we have relied on technology to solve complex problems and technology is a major economic force. There are estimates that the global AI market will reach US~$1.8 trillion in value by 2030. A McKinsey survey last year found that ‘nearly one-quarter of surveyed C-suite executives say they are personally using gen AI tools for work, and more than one-quarter of respondents from companies using AI say gen AI is already on their boards’ agendas.’

But where to begin?

It makes sense to start applying AI to low-value, low-risk tasks – to automate the mundane. This allows us to develop proof of concept and garner associated buy-in, develop basic competence, and gain further insight through implementation. The big picture idea here is that AI enables the procurement team to shift its expertise from an operational focus to a value-creating focus, heralding the next generation of procurement.

If you are looking for the low-hanging fruit then think of daily housekeeping activities, consider inflation reporting, benchmarking, contract and risk management, spend analytics and ESG measures as worthwhile starting points.

Perhaps one of the most obvious areas of application is sourcing, since generative AI is trained on large datasets and sourcing typically has plenty of data to work with, such as vendor information, contracts and purchase orders. The raw material is there for AI to generate RFPs and contracts.

It may be that it will take time for AI to fully develop an RFP or a contract, but it is entirely feasible that it will be able to complete a high proportion of such a task and continue to iteratively improve until it is virtually autonomous.

Figure 1: Typical low-risk, low-value applications of AI in procurement

AI is brilliant, but it is not human

There is some criticism of AI’s ability to discover insights in the same way that humans discover insights. But the criticism is somewhat unfair. It may be true that AI struggles to deal with humour or to recognise whimsy, but it can beat chess masters.

So, while the nuance or absurdity in humour might not pattern as well as AI might like, there are other patterns that AI can identify and excel at. Like a chess master, AI can see price patterns and even predict future fluctuations in prices and this capability can directly assist procurement teams actively manage inflationary pressures, as well as other factors such as the reliability of delivery and quality, and even financial stability.

The same applies to tasks such as benchmarking, where prices and indices can be tracked across multiple sources, or when historical datasets need to be integrated with new ones.

We will also see applications to support an organisation’s ESG agenda in areas such as assessing emissions, supporting sustainability-driven decisions, and bringing public source data to inform organisational ESG requirements.

What is abundantly clear, is that AI will transform procurement, just as it will every other aspect of an organisation. We see it as a natural next-generation evolution as we shift from an operational focus to a more strategic, value-creating focus.

Figure 2: An integrated picture of where AI adds value to procurement

As procurement and supply chain consultants, we see our role as being an independent, objective advisor. Firstly, we want to be sure clients understand the landscape. What types of AI, and what possible applications should be considered, typically make up our early discussions with clients. For many of us, there is a learning curve we didn’t even think of two years ago. Still, now we know, and now we have a duty to educate ourselves and our organisations.

Plus, the early numbers are very promising. We have multiple case studies of 40% productivity gains and 20% risk mitigation of revenue loss from good AI implementation.

Discussion about risk always takes place early in the piece too, and that is every dimension of risk, from financial to ethical and back again.

We also need to be cognisant of the value chain each organisation is in. What are your clients doing? What are your suppliers doing? The more integrated the data flowing through the value chain, the better for everyone.

However, as bold and brave as this new world may be, we must not let it overtake our objectives and the genuine, unique needs of each organisation. Working backwards from a needs analysis is always the best place to start.