Enterprise AI for Inventory Optimization

Introduction

A world leader in the production of welding and cutting equipment and consumables offers products across personal protective equipment, filler metals, cutting and welding tools, gas control equipment, and related accessories. With over 40 global brands and 20 manufacturing plants, the manufacturer aimed to improve inventory management to efficiently meet customer needs.

Challenges

The manufacturer used a Plan for Every Part (PFEP) approach to manage their inventory. This approach accounted for variables including cost of goods sold and order variability to determine replenishment strategies for each part in each location. The PFEP required time-consuming and manual data aggregation across multiple dashboards and spreadsheets, resulting in infrequent updates.

Part-specific replenishment strategies presented additional complexity since the manufacturer chose between demand-driven (safety stock) and reorder point-based replenishment depending on the product line and location. The manufacturer recognized the significant challenges PFEP presented and looked for a scalable alternative.

Solution

To address these challenges, the manufacturer partnered with C3 AI to implement an AI-driven solution, C3 AI Inventory Optimization. Within 12 weeks, the team deployed the application to scale inventory optimization and reduce excess inventory while maintaining sufficient availability.

Results

With C3 AI Inventory Optimization, the manufacturer identified a 26% overall inventory reduction opportunity while increasing material availability by 0.5%. With C3 AI Inventory Optimization, the manufacturer can rely on a unified view of their global inventory, optimize inventory for individual parts across locations, increase service levels, and lower working capital and inventory management costs.

About the Company

• $2+ billion annual revenue in 2022
• 140+ countries in operation
• $470 million global inventory value
• 9,000+ employees

Project Highlights

• 12 weeks from kickoff to completion of pre-production application
• Generated over 25 million simulations to determine ideal Safety Stock or Reorder Point at a SKU-location level
• Three years of historical data integrated, comprised of 5 million rows of data across 8 North America facilities for 2 product lines
• Built an extensible data model with 18 C3 AI logical objects
• Configured the C3 AI Inventory Optimization application user interface to mirror SME workflows and flow of materials from suppliers, facilities, and distribution centers
• Fully automated the calculation of new reorder parameters
• Generated over 2 thousand ML recommendations