Food and Grocery Face New Labor Challenges as COVID-19 Re-Shapes Distribution Centers

Artificial intelligence technology provides flexibility to adapt operations and systems as the grocery supply chain continues to evolve in the face of changing consumer shopping habits.

Ipopba Adobe Stock Ai Hands
ipopba - Adobe Stock

Grocery stores have faced increased demand and explosive growth in online orders ever since the pandemic lockdowns earlier this year. The e-commerce grocery surge has led to an investment boom by retailers in highly automated local fulfillment capabilities, but a very different situation is playing out in grocery and food distribution centers (DCs).

The DCs are dealing with temporary supply chain disruptions, staffing issues and health and safety concerns that have increased labor costs and exacerbated existing hiring and retention challenges. Big-ticket automation and new-age robotics solutions don’t solve these short-term issues.

On the other hand, artificial intelligence (AI) tools can address the current labor challenges and generate immediate double-digit productivity gains. Moreover, AI technology provides flexibility to adapt operations and systems as the grocery supply chain continues to evolve in the face of changing consumer shopping habits.

Improving worker safety while adapting to rising demand

Grocery distributors have been insulated from the worst economic fallout of the Coronavirus disease (COVID-19), with double-digit growth, but increased demand has led some grocery DCs to add shifts and workers. The pandemic has also created new operational challenges that directly impact labor productivity and efficiency, increasing costs per case shipped.

Lucas Systems Productivity GainsLucas Systems

DCs have taken various steps to improve employee safety and promote social distancing among workers, in line with local and national health department guidelines. They have increased cleaning and disinfecting within facilities (including of mobile devices and equipment) and many require workers to wear face coverings.

A number of other steps to enable physical distancing can make order selection less efficient. Some DCs have added shifts or workers to meet increased demand, in some cases reducing throughput per labor hour. Some are also staggering breaks and limiting employee cross-over between shifts. Others are dedicating a single worker per zone in higher density pick modules and others are converting aisles to one-way travel.

Many of these operational changes directly impact the DC’s existing RF- or voice-directed process flows and throw off labor standards based on previously defined standard work processes. Likewise, engineer-intensive slotting programs have been challenged to keep up with unexpected demand spikes and unforeseen product shortages. In short, the efforts to ship more volume while enhancing distancing have upended years of successful productivity initiatives.

Current technologies and robotic systems don’t address these challenges

Grocery DCs have been leaders in using technology to improve efficiency and minimize labor costs. They were among the pioneers to broadly adopt voice directed picking and they were also leaders in optimizing picking processes to reduce travel through lean engineering, slotting and engineered labor standards.

Despite those efforts, travel time still represents more than 30% of an order selector’s work day, and current picking systems don’t easily adapt to new distancing requirements. Likewise, traditional big-ticket automation (for de-palletizing, automated storage and retrieval, conveyors, sorters, palletizing and shipping) doesn’t make economic sense in today’s changing grocery market, where consumer demand is shifting to online ordering and local pickup and delivery. Highly automated DCs cost upwards of $100 million and take 1-2 years to build, putting them well out of reach for most companies.

New autonomous mobile robots (AMRs) are changing the face of e-commerce fulfillment operations, but current AMR solutions are designed primarily for each picking applications, so they are not optimized for case pick to pallet operations in the grocery DC.

AI is one emerging technology that shows tremendous promise to address the DC labor challenge without restructuring warehouse infrastructure or systems. AI is already being used for supply chain planning, but it is not yet widely used for DC process optimization and planning.

An AI approach to warehouse productivity

AI offers a non-obtrusive approach to adapt to new operational requirements. In particular, AI-based tools can be used to reduce selector travel by optimizing how work is organized and how selectors travel to complete their work assignments. This software-based approach for picking optimization can take account of one-way aisles and other process requirements, and can adapt to process and navigation changes.

In trials with grocery DCs, the technology has demonstrated travel reductions of up to 30% in case picking operations. That represents significant labor cost savings in labor-intensive order selection processes.

Intelligent work creation and pick sequencing

Current work creation and pick sequencing rules in most DCs do not optimize for selector travel. For example, warehouse management system-directed picking systems use simple volumetric rules to create 2-3 pallet assignments of work based on maximum cube per assignment (typically for a single store). Likewise, selector travel paths are not based on a minimum travel distance based on the ideal path in which to pick the items.

AI-based optimization

AI can get at the root of the problem by applying travel optimization algorithms alongside cubing and other pallet creation rules. For every two- or three-pallet work assignment, the algorithms determine which pallets can be picked together most efficiently in a unit of work (or “batch”). The tools also identify the shortest possible travel path and create optimal pick sequences for selectors to follow. Together, assignment and pick path optimization maximize pick density and reduce the total distance traveled per assignment.

In trials with several food and grocery DCs, AI-based work creation (batching) has demonstrated travel savings between 15-22%. Path optimization generated an additional 5-15% travel savings. This 20-30% travel savings translates to an 8-15% gain in order selection productivity (depending on how high a percentage of travel is in your current process).

A flexible approach to warehouse optimization

Using AI to reduce travel does not require any up-front investments in expensive new automation systems or massive process reengineering and warehouse restructuring. It is also a flexible solution that can adapt to changing order profiles as DCs start to ship more products to online-only dark stores or micro-fulfillment centers.

Machine learning – a particular form of AI that uses learning algorithms to find relationships in large Internet of Things datasets – also shows promise for reducing travel by improving the responsiveness of slotting systems. Rather than spending weeks of engineering time to adjust your slotting models, Machine learning tools can analyze current and historical product, order and picking data to provide managers with immediate slotting recommendations,


Increased demand and new safety concerns have collided to create a range of new productivity challenges for food and grocery DCs. Even before COVID-19 took hold, many of these facilities were facing long-term hiring and retention challenges. In addition, changes in how and where consumers buy groceries will have ripple effects in food and grocery distribution operations for years to come.

While robotic picking systems have gotten most of the attention as a cure for DC labor issues, AI technology has immediate advantages for food and grocery DCs. It represents a non-obtrusive solution for improving efficiency by reducing travel, while taking account of current distancing and staffing measures meant to keep workers safe and healthy.