The client is a multinational chemical manufacturing company that has more than one million MRO spare parts in its warehouse. It was using an enterprise-wide inventory management solution (MDRM) to manage the catalogue, but an ineffective in-built search functionality led to the creation of duplicate items in the catalogue as the material engineers would create new purchase requisitions upon unsuccessful search. The existing solution separately provided a rule-based mechanism to identify these duplicates/similar parts to keep the catalogue accurate but the false positive rate was quite high. The material engineers on average approved less than 1% of the recommendations upon review. Over time, the lack of visibility into duplicate items and ineffective detection resulted in the accumulation of unused inventory items, high inventory levels and suboptimal utilization of Capex. It also resulted in inefficient sourcing of parts and price deviations for similar parts. Lastly, the client had to bear the risk of plant shutdown in times of urgency due to a lack of ability to find correct substitutable or similar spare parts.

‍

To alleviate these issues, the client collaborated with Soothsayer to build a comprehensive AI-driven duplicate detection framework well-integrated with their existing business workflows.