Do procurement teams already use AI?

Yes, but often not through a formal procurement program. The Microsoft Work Trend Index 2024 says 75% of knowledge workers use AI, 78% bring their own AI tools to work, and only 39% have received company AI training.

That pattern matters for procurement because the work is document-heavy and decision-heavy. A category manager can paste a supplier quote into a public tool. A buyer can ask for a contract summary before a meeting. An analyst can ask for a risk table from a spend extract. None of those actions requires a central program to exist first.

The question is not whether AI has entered the team. It probably has. The question is whether the team has agreed what counts as acceptable use, what data may be pasted, and which outputs require verification before they shape a supplier decision.

Bring-your-own AI creates uneven practice. One person may use AI only for drafting email. Another may use it to rank bids. Another may avoid it entirely. Without a shared method, the team gets different levels of review on similar work.

Low formal training leaves control gaps. If fewer than half of knowledge workers using AI have received company AI training, managers should assume that some use is informal. That is not a reason to ban it by default; it is a reason to give clear rules for data handling, source checking, and final human accountability.

Procurement needs task-level rules. A generic AI policy is not enough for a quote comparison, a supplier-risk register, or an MRO catalog merge. The useful question is concrete: for this task, which fields can AI draft, which fields must be checked, and where does the final decision sit?