AI
Smarter ways to scope, shortlist and select suppliers Choosing suppliers has long been a labour-intensive task. Manual searches, fragmented data and inconsistent comparisons slow down the process. Machine learning simplifies this by automating intake, improving scoping and creating shortlists of vetted suppliers.
Globality’ s platform, for example, introduces agentic intake, where autonomous software agents collect contextual information through conversation rather than static forms. This helps scope complex requests more accurately and feeds a richer data set into the supplier recommendation process.
“ When it’ s time to recommend suppliers,” Keith says,“ we apply agents that build and refine a supplier registry over time. These agents rate suppliers based on categories, commodity codes and performance data, before presenting an ideal mix of potential partners.”
Machine learning models then assess those partners across price, quality and sustainability metrics using clustering techniques – methods that group data points based on shared characteristics. Once a shortlist emerges, analytics powered by machine learning takes over again to evaluate proposals, uncover risk and apply game theory, a mathematical framework used to model strategic interactions, helping buyers optimise negotiation outcomes.
The result is a procurement function able to extract more value, work more efficiently and strengthen supplier relationships without increasing overheads.
“Procurement professionals were among the first to experiment with generative AI”
Keith McFarlane, Chief Technology Officer, Globality
Adoption hurdles and the need to build trust While the benefits of machine learning in procurement are visible, barriers remain. Chief among them is internal resistance.
“ Anxiety around AI – whether it’ s job displacement, data security or fear of wasted investment – often breeds resistance to change,” says Keith.
High-profile AI failures have not helped. A report from the Massachusetts Institute of Technology( MIT) outlines common problems, including building too many tools in house, underestimating how complex integration can be and misjudging where value actually lies.
Keith insists that education is the way forward. Procurement teams must learn that machine learning agents do not replace people, but expand capacity. With routine analysis and repetitive work offloaded, professionals can focus on higher-value activities and manage more spend categories.
One case in point is Heineken, which is using machine learning not in sourcing but in supply chain planning.
100 November 2025