How Baker Hughes Uses AI to Optimize Sourcing Across 800+ Global Sites

Baker Hughes operates at a scale that makes supply chain excellence both essential and unusually difficult. Serving the global energy sector across 120 countries with a product catalog spanning hundreds of thousands of parts, the company’s procurement teams carry enormous responsibility for cost control, supplier resilience, and delivery continuity. At that scale, the quality of sourcing decisions matters enormously — and the quality of data underlying those decisions matters even more.

Before deploying AI-driven sourcing, Baker Hughes procurement teams operated in an environment of fragmented information. Critical data was distributed across disconnected ERP systems, spreadsheets, and external feeds. Buyers frequently worked with stale insights: by the time an analysis was completed and actioned, purchase orders had already passed their cancellation window or market conditions had shifted. The problem was compounded by structural inconsistencies — different ERP systems occasionally assigned identical part numbers to different items, and foreign exchange fluctuations created apparent savings that dissolved under scrutiny. Ground-level teams sometimes sent purchase orders as placeholders to hit order cycle targets, obscuring the true picture of demand.

In 2019, Baker Hughes partnered with C3.ai to address this systematically. The initial deployment of C3 AI Sourcing Optimization began with a three-site pilot before expanding to ten sites and ultimately to enterprise-wide deployment across more than 800 facilities. The system integrates data from five enterprise sources — over 30 million records in total — and presents sourcing opportunities to buyers as complete evidence packages: total landed cost breakdowns, similar-part rationale for leveraging supplier negotiations, and index-based pricing signals. A human-in-the-loop feedback mechanism allows buyers to accept or reject recommendations, continuously improving the model with institutional knowledge. Mobile access via iOS and Android ensures the system is available wherever buyers operate.

The most striking result came in Accounts Payable-adjacent sourcing, where the team achieved a 5–7% acceptance rate on AI-identified opportunities — a figure that, applied to a cost-of-goods-sold base in the high seven- to low eight-figure annual range, translates to millions in realized savings. By 2025, the program covers 900,000 SKUs across 800+ program facilities with 200+ active users. The team runs weekly business reviews to capture savings in real time and sustain adoption momentum.

Looking ahead, Baker Hughes has embedded AI not as a reporting layer but as an active decision system. The ability to monitor live tariff feeds and instantly model the total landed cost of an order as new executive orders are announced gives procurement teams genuine market responsiveness — a capability that has become critically valuable in a period of significant supply chain volatility. The program positions Baker Hughes to extend AI-driven optimization beyond sourcing into adjacent procurement processes.