The client is a USA-based multi-brand e-commerce retailer operating seven regional fulfillment centers. As order volumes and SKU assortment expanded, popular products regularly sold out in one region while the same items accumulated excess stock in another.
The retailer needed inventory optimization software that could connect warehouse availability, order demand, inbound supply, transfer lead times, fulfillment SLAs, and shipping economics. Planners needed earlier visibility into regional stock risks to prevent split shipments, cross-zone deliveries, or higher safety stock.
Computools delivered StockBalance AI, an AI inventory management platform integrated with the client’s existing WMS and OMS environments. Within five months of phased adoption, the platform reduced split shipments by 21%, lowered safety stock by 14%, reduced cross-zone parcel costs by 11%, and decreased regional stockouts by 24%.
The client is a USA-based e-commerce retailer selling home, lifestyle, consumer electronics, and accessory products through its own storefronts and major online marketplaces. The company fulfilled orders through seven regional distribution centers across the United States. Its distributed warehouse footprint helped keep products closer to customers, but regional demand volatility made inventory positioning harder to manage as the assortment grew.
Existing warehouse inventory management software and OMS tools supported warehouse transactions, stock movements, and order routing. However, they did not continuously determine how much inventory each facility should hold by SKU, region, demand velocity, and service-level target.
Leadership needed a fulfillment optimization platform that could improve inventory placement without replacing the systems already running warehouse execution and order operations.
The main challenge was the lack of visibility into future SKU-location demand. The retailer had enough warehouse, order, and supply data to detect imbalances, but planning teams still relied on spreadsheet exports, historical averages, and weekly review cycles.
The project required warehouse management software development experience across WMS and OMS integration, warehouse operations, SKU-location planning, fulfillment economics, and phased rollout.
Computools designed StockBalance AI as an inventory intelligence and allocation layer for the client’s multi-warehouse fulfillment network. The platform connected WMS, OMS, procurement, carrier, and planning data into one decision environment.
The system ingested inventory positions, sales orders, order-line velocity, backorders, inbound purchase orders, inter-warehouse transfers, promotions, regional demand patterns, fulfillment lead times, and parcel shipping costs.
The forecasting layer worked as demand forecasting software, generating SKU-location demand forecasts and projected inventory positions for each fulfillment center.
Computools applied AI development to support forecast generation, safety-stock calculation, stockout risk scoring, allocation logic, and recommendation explanations.
Warehouse stockout prediction logic identified future shortages before they affected customer orders. The allocation engine compared projected demand, available stock, inbound inventory, transfer lead time, and fulfillment cost to identify where inventory should move.
The inventory rebalancing software generated recommended transfers with origin warehouse, destination warehouse, quantity, expected stockout risk, projected days of supply, estimated transportation impact, and confidence level. Planners could approve, adjust, or reject each recommendation.
StockBalance AI was piloted across two fulfillment centers, then expanded after forecast quality, planner adoption, transfer impact, and fulfillment KPIs were validated.
Within five months of phased adoption, the client achieved measurable improvements across inventory placement, fulfillment cost, stock availability, and planning speed:
Warehouse stock optimization helped planners position inventory by SKU and region, reducing manual spreadsheet work. Teams could separate network-wide shortages from location-specific imbalances and act before availability issues affected fulfillment.
Earlier rebalancing recommendations improved supply chain inventory optimization by linking stock movements to demand forecasts, service levels, transfer costs, and parcel shipping economics.
The client selected Computools because the project required warehouse workflow analysis, inventory data modeling, AI forecasting logic, WMS and OMS integration, planner UX, and phased rollout support.
Computools turned fragmented warehouse, order, inbound supply, transfer, and carrier data into a practical decision model for allocation, rebalancing, and stockout prevention.
Delivery priorities were tied to measurable KPIs: split-shipment rate, regional stock availability, safety stock, cross-zone parcel cost, transfer planning speed, and fulfillment SLA performance.
The team covered business analysis, CX and UX design, architecture, AI logic, integrations, QA, rollout validation, and post-launch optimization.
The design focused on exception-driven planning, SKU-location visibility, transfer recommendations, stockout risk, and network-level fulfillment economics.
Defining warehouse planning needs, fulfillment risks, inventory transfer workflows, stock availability, and network cost visibility.
Structuring the platform around network inventory visibility, SKU-location risk, transfer recommendations, and fulfillment economics.
Designing low-fidelity layouts for network inventory views, stockout-risk queues, SKU-location analysis, transfer approvals, safety-stock rules, fulfillment-cost comparison, and executive dashboards.
The interface gave planners one place to review inventory risk, compare warehouse positions, approve transfers, and track fulfillment impact across the network.
TypeScript
TypeScript supported consistent frontend and backend development across inventory views, allocation workflows, approval logic, and analytics interfaces.
Next.js and React
Next.js- and React-powered responsive planning and analytics interfaces for inventory, fulfillment, transportation, and executive teams.
Node.js and NestJS
Node.js and NestJS handled application logic, recommendation workflows, approvals, notifications, user permissions, and integration orchestration.
PostgreSQL
PostgreSQL stored products, fulfillment centers, inventory snapshots, transfer recommendations, approval history, planning parameters, forecasts, and KPI history.
Python and FastAPI
Python and FastAPI powered regional forecasts, stockout-risk models, safety-stock calculation, allocation scoring, and rebalancing services.
AWS, Docker, and Terraform
AWS, Docker, and Terraform supported cloud deployment, environment consistency, infrastructure automation, monitoring, and scalable processing.
OpenAI and Claude integrations
Controlled LLM integrations supported natural-language explanations for inventory risks, transfer recommendations, and projected business impact using structured platform data and predefined analytical workflows.
Computools used Scrum with two-week iterations, stakeholder demos, and KPI-based acceptance criteria. Each sprint connected a platform capability to a measurable operational outcome: split shipments, stock availability, safety stock, transfer planning speed, or fulfillment cost.
The team started with workflow mapping, WMS and OMS data review, and SKU-location inventory modeling. Forecasting, stockout-risk detection, and rebalancing recommendations were introduced after data completeness and inventory consistency reached agreed thresholds.
Two fulfillment centers were selected for the pilot. The locations represented different regional demand patterns, allowing the team to validate forecast quality, recommendation acceptance, transfer effectiveness, and fulfillment impact before network-wide rollout.
Before StockBalance AI, we knew how much inventory we owned, but we could not always tell whether it was in the right place. Computools helped us see inventory risk across the network early enough to act on it. We reduced split shipments, lowered buffer stock, and gave planners a clearer way to manage regional availability.
Computools | 341 Raven Cir, Camden Wyoming, Delaware 19934, US
