Quiet Wins, Hard Limits: A Procurement View of AI

A risk-averse procurement function is quietly capturing AI savings in contract review and negotiation while resisting hype, guarding its data in contracts, and absorbing a hardware cost shock driven by the AI boom itself.

One IT procurement contract manager in the oil and gas sector describes AI as a genuine but unmeasured force in his work. AI tools analyze vendor redlines, simulate negotiations, and benchmark markets. He believes these tools materially lifted cost-reduction numbers and expanded each negotiator's workload capacity, but his team tracks only cost reduction and avoidance, so it cannot attribute value to any single tool. He is skeptical of automated negotiation bots, which fit commodities like paper but not the assembled IT products he buys, and he notes his firm deliberately lets others be the guinea pig. On governance, he describes specific contract language: vendors may train models on company data only for that company's isolated model, never for competitors, and the firm claims ownership of AI outputs. Vendors resist this. He flags data privacy and risk mitigation as the main constraints, with uneven data quality a secondary worry. He sees the AI infrastructure boom as a structural cost shock, with memory and hardware prices spiking and likely to stabilize high rather than fall. Despite his cautious employer, he personally argues that large organizations are too slow and should run more pilots and proofs of concept, expecting process optimization, savings, and analyst capacity gains. His parting wish is a clear forecast of how automated work and life will become, so firms can plan now.