Beena Ammanath, Executive Director, Predix Data Science
General Electric spends approximately $40 billion in direct materials sourcing and has hundreds of ERP systems across its different business units. As the systems and business units are fragmented and operating in silos, they often have multiple contracts with the same supplier for delivering same or similar products. Supplier partners are restricted to a particular GE business they cater to and do not have visibility into the needs of other GE business units that might need the same or similar product, leading to multiple contracts, negotiations and a duplication of effort for both sides. There was a significant opportunity to bring all the businesses and systems together and identify opportunities to unify sourcing efforts, which would benefit both GE and its supplier partners in multiple ways.
With the help of machine learning and analytics, this solution analyzed entire GE industrial direct materials sourcing, drove a new level of transparency and visibility to all GE business units, the parts they need, price points, to enable better negotiation and transparency with GE supplier partners. It also helped supplier partners to work with multiple GE business units for same or similar parts. Examples of solutions implemented:
- Image analytics of parts for design similarity, when the part name/descriptions are not matching, to identify similar parts being sourced from different vendors
- Enterprise-wide search for all sourcing teams across GE, to search for part/supplier and the price point being offered in the past, for efficient future purchase decision leveraging Natural Language Processing
- Identification of suppliers providing same parts across contracts and businesses through big data engineering and machine learning algorithms
GE: Estimated productivity gains of up to several million dollars through analytics initiatives around supplier mastering, parts harmonization and payment terms mastering.
Supplier Partners: Transparency and visibility to other GE business unit needs via semantic search tool and enabled scale economy through additional orders for the same part.