How Baker Hughes Uses C3 AI to Optimize Sourcing Across 800+ Sites

Baker Hughes operates one of the most geographically dispersed supply chains in the energy sector—800-plus facilities across more than 120 countries, managing hundreds of thousands of unique parts and suppliers. That scale creates a data problem: critical sourcing information sits in disconnected ERP systems, spreadsheets, and supplier databases that don’t talk to each other. By the time a buyer collects and reconciles the data needed to evaluate a purchase decision, market conditions have often changed, purchase orders have passed their cancellation window, or the analysis is obsolete.

The procurement pain points were systemic. Different ERP systems occasionally shared identical part numbers for entirely different items, creating dangerous confusion at scale. Foreign exchange fluctuations generated phantom savings figures that made decisions look better than they were. And on the ground, teams sent purchase orders as placeholders to hit order cycle targets—further distorting the data landscape that analysts relied on. Buyers with deep industry expertise were, in effect, buried under bad data.

In 2019, Baker Hughes partnered with C3.ai to address this at the foundation. C3 AI Sourcing Optimization ingests data from 30+ million records across five enterprise systems, integrates live tariff feeds, and surfaces sourcing opportunities with a full evidence package: total landed cost breakdowns including logistics, similar-part comparisons for supplier leverage, and real-time index-based pricing signals. Buyers see opportunities ranked by value and can accept or reject recommendations in a human-in-the-loop workflow—each action training the system’s future recommendations with institutional knowledge. The platform is mobile-enabled on iOS and Android.

What began as a pilot across three sites scaled to ten, then to enterprise-wide deployment across 800-plus facilities. The business model for savings is grounded in high-volume product lines: focusing on those areas, the team achieved a 5–7% acceptance rate on identified opportunities—significant given that cost of goods sold typically runs in the high seven- to low eight-figure range annually. The platform realized millions of dollars in validated economic benefit in CY2025, with savings captured in weekly business reviews through live buyer feedback.

The real advancement is speed of insight. Baker Hughes buyers can now monitor live tariff data and immediately calculate total landed cost as new executive orders or market shifts are announced—turning what was a days-long reconciliation process into a near-real-time decision. With 200+ active users and a foundation of 30M+ integrated records, the platform has shifted from a dashboard into an enterprise-scale sourcing intelligence layer.