Pick Path and Wave Optimization

Warehouse picking remains the most labor-intensive activity in fulfillment operations, with order picking accounting for approximately 55% of total warehouse operating costs according to industry benchmarks reported by Productiv in 2025. Labor itself represents 50% to 70% of total warehouse operating expenses, as noted by ExploreWMS, making even marginal efficiency gains in picking operations financially significant. The 2025 Warehousing and Fulfillment Costs and Pricing Survey, which gathered data from more than 600 warehouses, found that picks per hour per staff member increased from 64.23 in 2024 to 102.33 in 2025, suggesting that process optimization and automation are accelerating across the industry.

The core challenge lies in the compounding inefficiencies of manual pick path planning and static wave release strategies. Workers in poorly optimized facilities may walk more than 10 miles per day, and travel often accounts for half of all labor time in order picking, according to Lucas Systems. Poorly constructed waves that are too large, too small, or misaligned with available labor add administrative overhead without delivering efficiency benefits, as documented in a 2024 Optioryx analysis of warehouse picking strategies. These inefficiencies multiply in omnichannel environments where distribution centers must simultaneously process e-commerce, wholesale, and store replenishment orders with different service-level requirements.

The financial pressure is intensifying as warehouse profit margins tighten. The 2024 Warehousing and Fulfillment Costs and Pricing Survey reported that corporate profit margins decreased from 10.49% in 2023 to 9.37% in 2024, underscoring the urgency for operational cost reduction through technology-driven optimization.

AI-driven pick path and wave optimization operates across three interconnected layers: intelligent wave planning, dynamic route generation, and predictive slotting. At the wave planning level, machine learning models cluster incoming orders into waves based on delivery windows, carrier cutoff times, order characteristics, and real-time labor availability. A 2024 simulation study published in Scientific Reports (Nature) found that consolidating orders into waves of optimized sizes achieved up to a fourfold reduction in total pick distance, though the authors noted that effectiveness is highly sensitive to wave size and order clustering parameters.

For dynamic pick path generation, AI systems solve variants of the Traveling Salesman Problem and the Shortest Path Problem to calculate optimal travel sequences for each picker. These algorithms consider warehouse layout geometry, aisle directionality, current congestion levels, product locations, and picking method (batch, zone, cluster, or discrete). The systems ingest order, inventory, and location data from warehouse management systems and apply real-time optimization algorithms to create work assignments that minimize empty travel between picks. Grid Dynamics reported in a published case study that an explainable AI system using Random Forest models to dynamically reassign item locations and optimize pick paths reduced average order picking time by 23% across multiple warehouse zones.

Predictive slotting intelligence forms the third layer, using demand velocity data, SKU affinity analysis, and seasonal patterns to continuously adjust product placement. This reduces the average distance between frequently co-ordered items and balances workload across warehouse zones. Integration with existing warehouse management and enterprise resource planning systems remains the primary implementation challenge, as MIT researchers cited in a 2025 Mecalux-MIT study identified poor data quality, legacy system constraints, limited technical expertise, and budget limitations as the main barriers preventing full AI value realization in warehouse environments.

Organizations should note that these systems require clean, granular operational data to function effectively, and results vary significantly based on warehouse layout complexity, SKU count, and order profile diversity. Facilities with fewer than 5,000 SKUs or simple single-channel order profiles may see diminished returns relative to implementation costs.

DHL Supply Chain, the global logistics division of Deutsche Post DHL Group, has deployed AI-driven optimization across multiple warehouse formats. In one documented implementation, DHL used a simulation-based optimization tool called IDEA to improve cluster picking in medium-size warehouses, reducing heavy aisle congestion (four or more carts in one aisle) from 28% of the time to 18% while reducing the number of pickers needed to maintain throughput. In a separate warehouse layout optimization project for a European footwear retailer, DHL partnered with analytics firm Logio and achieved a 22% reduction in picker walking distance and an estimated 8% increase in picking productivity through data-driven slotting and layout redesign. DHL has reported overall results including up to 50% reduction in warehouse employee travel distance and 30% productivity increases in order picking across facilities using AI-driven routing and resource allocation.

In the manufacturing and distribution sector, Southwire, a leading North American wire and cable manufacturer supplying major retail chains, implemented the Lucas Systems Dynamic Work Optimization solution to supplement the company's warehouse management system. The deployment achieved a 30% to 50% increase in e-commerce picking productivity, a 90% reduction in training time for new pickers, and a twofold increase in lines picked per hour, according to Lucas Systems. A separate pet supply retailer using the same platform reported a 33% increase in throughput. These results were achieved without changes to warehouse layout or the addition of robotics, relying solely on AI-based batching and path optimization layered on top of existing infrastructure.

The market for AI-driven pick path and wave optimization spans enterprise warehouse management system providers, specialized warehouse execution system vendors, and niche optimization solution firms. The Warehouse Management System market exceeded $3 billion in 2024 with a five-year compound annual growth rate forecast in double digits, according to the 2025 Gartner Magic Quadrant for Warehouse Management Systems. The 2025 Gartner report evaluated more than 80 WMS providers, with 17 meeting documented inclusion criteria. Manhattan Associates was positioned as a Leader for the 17th consecutive year, receiving the highest scores for Level 3 through Level 5 complexity warehouse operations in the accompanying Gartner Critical Capabilities report.

Selection criteria for pick path and wave optimization solutions should include the depth of AI and machine learning capabilities for dynamic task orchestration, support for both wave-based and waveless (continuous order streaming) fulfillment models, integration architecture with existing warehouse management and enterprise resource planning systems, robotics and automation orchestration capabilities, and the vendor approach to data quality requirements and implementation support. Cloud-native architectures are increasingly preferred for scalability and continuous feature delivery.

• Manhattan Associates -- cloud-native warehouse management with machine learning-driven order streaming, real-time pick path optimization, integrated slotting, and warehouse execution system capabilities for Level 3 through Level 5 complexity operations
• Blue Yonder -- enterprise warehouse management with AI-driven prescriptive task interleaving, dynamic slotting, robotics orchestration hub, and Microsoft Azure-based cloud architecture
• Körber Supply Chain -- adaptable warehouse management portfolio with voice-directed picking, multi-client support, and configurable picking optimization for mid-to-large distribution centers
• Lucas Systems -- specialized AI-based warehouse optimization with Dynamic Work Optimization for intelligent batching, pick path optimization, and voice-directed workflows that integrate with existing warehouse management systems
• Softeon -- integrated warehouse management and execution system with rules-based wave and waveless order release, picking sub-system orchestration, and simulation-based labor planning
• SAP SE (Extended Warehouse Management) -- enterprise warehouse management integrated with broader ERP and supply chain management, supporting complex multi-site distribution operations
• Dematic (KION Group) -- warehouse automation and software solutions combining goods-to-person systems with AI-driven orchestration for high-volume fulfillment environments