This Philippine retail chain operated 78 stores across Metro Manila, Cebu, and Davao, spanning supermarket, convenience, and home goods formats with over 45,000 SKUs. Inventory management was the chain's most persistent operational challenge. Stockout rates averaged 12% of SKUs at any given time — meaning customers regularly could not find what they came to buy — while simultaneously carrying PHP 340 million in excess and slow-moving inventory across the network.
The root cause was a demand forecasting system based on simple historical averages and manual adjustments by category managers. Philippine retail demand was influenced by an unusually complex set of factors: payday cycles (the 15th and 30th of each month drove dramatic spending spikes), religious festivals across Catholic and Muslim communities, typhoon seasons that shifted buying patterns toward emergency supplies, and significant regional taste differences between Luzon, Visayas, and Mindanao.
Category managers made replenishment decisions for their product categories using spreadsheets, personal experience, and supplier relationship dynamics rather than data-driven demand signals. This created a cycle of alternating stockouts and overstocking, with an estimated PHP 280 million in annual lost sales from stockouts and PHP 95 million in markdown losses from excess inventory.
Pertama Partners conducted an AI Readiness Audit analyzing two years of point-of-sale data across all 78 stores, supplier lead time records, promotional calendars, and external data sources including weather forecasts, payday calendars, and local event schedules. The audit revealed that demand patterns were highly predictable when the right signals were combined — the problem was that no human analyst could process the volume and interaction effects of 45,000 SKUs across 78 stores with dozens of influencing factors.
Our AI Pilot Program deployed a demand forecasting and inventory optimization system at 15 stores spanning all three formats. The forecasting model generated SKU-store-day level demand predictions incorporating payday effects, weather sensitivity by product category, promotional lift modeling, local event impacts, and cross-product substitution patterns. The optimization engine translated these forecasts into automated replenishment orders, dynamic safety stock levels, and markdown recommendations for slow-moving items.
The AI Transformation Program scaled the system to all 78 stores and integrated it directly with the chain's warehouse management and supplier ordering systems. Team Training equipped category managers to shift from manual replenishment planning to exception-based management — reviewing and approving AI recommendations rather than calculating orders from scratch. Executive Training helped the leadership team understand how AI-driven inventory optimization could fund the chain's expansion plan by freeing up working capital locked in excess stock.
"Our category managers are brilliant merchants, but asking them to forecast demand for 45,000 products across 78 stores is asking the impossible. Pertama Partners gave them an AI partner that handles the math so they can focus on the merchandising."— Carlo Reyes, Chief Supply Chain Officer
Every transformation starts with a conversation. Let's discuss your challenges and opportunities.
Discuss Your Challenge