A truck delivers fresh produce to a supermarket and heads toward its next assignment. The produce has been received, but the logistics transaction is not complete. There are empty crates in the store that belong somewhere else, must be cleaned before being reused, and need to return to circulation before the next shipping cycle. In a disposable packaging model, much of that work ended with a recycling or waste stream. In a reusable model, it becomes an operating obligation with its own assets, service levels, and economics.

Reusable pallets, crates, and consumer containers are commonly discussed in environmental terms. That is understandable, given the potential to avoid single-use material. But for logistics executives, a more interesting development is taking place: the return trip itself is becoming a product. Someone must forecast the containers, retrieve them, inspect or sanitize them, reposition them, and have the right quantity available at the next point of use. A reusable package is not an environmental benefit simply because it was designed to be reused. It has to come back, in usable condition, often enough to justify the system around it.

Whole Foods makes the recovery network visible

In November 2026, IFCO described an expansion of its program with Whole Foods Market. The initiative began in 2022, initially replacing disposable corrugated cartons for products such as root vegetables with reusable plastic containers. IFCO said the collaboration had avoided nearly four million pounds of cardboard waste since inception and that use of the containers had increased substantially over the period.

Those are supplier-reported environmental and scale figures, but they point to a practical network-design question. A crate arriving with produce must eventually be separated from the merchandise, consolidated, collected, inspected, cleaned, and released for another trip. It also has to arrive back in circulation in the right geography and at the right time. A container sitting in a stockroom is an unavailable transportation asset, even if it has not physically left the supply chain.

This is why a returnable-packaging program cannot be managed solely as a procurement substitution—plastic instead of cardboard, one unit price instead of another. Store operations, food safety, transportation planning, distribution, and the pooling provider all influence whether the loop works. The more locations a program serves, the more important container visibility, loss prevention, route density, and balancing of empty inventory become.

Ocado turns the delivery route into a collection route

The e-commerce version is particularly revealing. Ocado Retail and Amcor’s UK refill project lets consumers receive staples such as rice, pasta, and laundry liquids in purpose-designed reusable containers, then hand the empties back to a delivery driver during a later visit. Amcor supplies the containers; GoUnpackaged helps operate the collection, pooling, and wash arrangements. In effect, an existing final-mile delivery system acquires a reverse flow of physical assets.

That is an attractive model because the vehicle is already visiting the customer. A return can potentially avoid a separate collection trip, although it still requires space on the vehicle, driver handling, sorting, cleaning, and an effective process for moving the recovered container back to a filler. The efficiency of the existing route helps, but does not make the reverse trip costless.

GoUnpackaged reports that more than 90% of dry-goods vessels sold over a twelve-month period were returned. That is an encouraging recovery metric, not a guarantee that each individual container completes a commercially sufficient number of cycles. Amcor and its project partners have discussed an estimated service life of 50 to 100 refills, with continued operation intended to establish the actual figure. There is a significant difference between engineering durability, return rate, and observed lifetime trips. The financial model needs all three.

A container that survives 75 wash cycles but routinely goes missing after eight customer deliveries will not achieve the expected cost per use. Conversely, a high return rate will not rescue an operation whose empty assets need expensive repositioning or whose cleaning capacity constrains distribution. The limiting factor could be consumer behavior, transportation, sanitation, or simply inventory management.

Iperal shows how returnable packaging affects automation

Packaging is also becoming part of automation architecture. In northern Italy, supermarket group Iperal worked with Tosca to introduce standardized reusable crates for its new automated fresh-food center in Giussano. Tosca reports that the consistent crate footprints and physical quality have supported steadier handling, more stable pallets, and fewer line interruptions than the prior mix of material formats.

The business implication is broader than the elimination of waste. Automated handling systems are sensitive to dimensions, stability, alignment, and repeatability. Standardized returnable containers can reduce variation at precisely the point where variation becomes expensive. But the same standardization also imposes an obligation: suppliers and stores must manage the containers as circulating assets, not as expendable packaging.

This is where packaging decisions meet material-handling engineering. The relevant question is not only the price of the crate; it is whether that crate reduces damage, manual touches, stops, and variability across the entire operating loop. The answer will depend on the facility, product mix, and recovery network. The claims in a supplier case study are useful starting evidence, not a substitute for a customer’s own cost and throughput model.

The pool is an operating company, not a packaging catalog

CHEP’s pooled-pallet model makes the network responsibility explicit. Customers obtain access to pallets without purchasing and independently managing every unit. CHEP organizes retrieval, inspection, repair, and reuse. Tosca provides comparable lifecycle services for reusable containers. These companies are not simply selling physical items; they are operating asset networks across multiple producers, retailers, distribution locations, and service centers.

That changes how suppliers should be evaluated. Unit purchase price gives way to a service equation that includes asset availability, recovery speed, hygiene, damage, billing accuracy, traceability, and performance during seasonal peaks. In a geographically dense network, return movements can often be combined with existing transportation flows. In a sparse network, repositioning costs and empty miles can erode the benefit quickly.

The economics can be reduced to a simple illustration. If a container costs $30 and completes 60 usable trips, the acquisition component is $0.50 per trip before washing, transport, administration, loss, repairs, and financing. If the same unit completes only 12 trips, that component rises to $2.50. Neither scenario is an actual customer result; both show why utilization matters much more than a favorable initial unit price.

The environmental calculation requires similar discipline. Life-cycle assessments must account for manufacturing, cleaning, loss, and transport as well as displaced single-use material. GoUnpackaged reports favorable results for certain Ocado categories under its assessed conditions. It would be a mistake to extrapolate that conclusion to every product and geography without testing the assumptions.

The next competitive advantage is control of the loop

Reusable packaging will not be suitable for every product, customer, or distribution pattern. Food safety standards, asset loss, reverse-mile density, and operational complexity will sometimes favor other formats. The discipline is to evaluate the complete cycle rather than to treat reuse as inherently superior.

What is changing is the commercial nature of the asset. The package is no longer finished when its contents arrive. Its next trip must be planned, and its performance must be measured. That makes the reverse flow a logistics service in its own right, with value increasingly determined by the intelligence and execution of the network. The winning innovation may not be a different crate. It may be a better way to get the same crate back.

The post The Return Trip Is Becoming a Logistics Product appeared first on Logistics Viewpoints.

Warehouse automation is often sold with a deceptively simple equation: invest in technology, reduce labor, increase throughput, and calculate the payback period. The reality is considerably more complicated. As warehouse operators spend more on robotics, automated storage and retrieval systems (AS/RS), autonomous mobile robots (AMRs), conveyors, sortation, and increasingly sophisticated software, the central investment question is changing. It is no longer simply, “How quickly will this system pay for itself?” A better question is: “What will this operation cost, constrain, and enable over the next five to ten years?” That distinction matters because some of the most valuable benefits of automation do not appear neatly in a labor-savings calculation.

The Automation Business Case Is Getting Bigger

Investment momentum remains strong. The 2026 MHI Annual Industry Report, produced by MHI and Deloitte, found that 56 percent of supply chain leaders were increasing technology and innovation investments. More than half, 52 percent, planned to spend over $1 million, while 17 percent expected to spend more than $10 million.

Warehouse automation competes for those investment dollars, but the range of potential projects is enormous. A relatively contained AMR deployment is economically very different from installing a high-density AS/RS, automated sortation, new picking stations, and the software required to orchestrate the entire operation. That makes comparisons based solely on capital expenditure misleading.

The real cost of automation includes equipment, software, integration, implementation, site preparation, training, maintenance, support, and the operational disruption associated with deployment. It can also include WMS modifications, network infrastructure, controls integration, safety systems, spare parts, and ongoing software subscriptions. In other words, purchase price is not total cost of ownership.

Labor Savings Matter, but They Are Only the Beginning

Labor remains one of the clearest economic arguments for automation, particularly in repetitive activities such as transportation, picking, replenishment, and sortation. But the most effective automation projects do not necessarily eliminate people. Instead, they change what those people can accomplish. Consider DHL Supply Chain and its deployment of Locus Robotics AMRs at a Tennessee life sciences operation. According to a Locus Robotics case study, productivity increased from an initial 78 units per hour to more than 150 units per hour. DHL was able to process more than twice the order volume while maintaining the same workforce headcount, while order cycle time fell by 50 percent.

That illustrates an important distinction in automation economics. The financial return does not have to come from replacing ten employees. It may come from enabling those employees to support twice the volume, reducing overtime, limiting seasonal hiring requirements, shortening training, or delaying the need to add another shift. For operations experiencing rapid growth, labor avoidance can be just as economically significant as labor reduction.

The Forgotten Variable: Space

Real estate can dramatically alter the automation equation. Traditional ROI calculations tend to focus on direct operational savings. But AS/RS and other high-density technologies can increase the amount of inventory that fits within an existing building. That creates an economic benefit that is easy to overlook.

Imagine a distribution center approaching capacity. Management could lease additional warehouse space, expand the existing facility, relocate, or redesign storage using automation. If automation allows the company to remain in the existing building for another five years, the avoided real estate, relocation, duplicated inventory, and labor expenses belong in the business case. This is why a system with a higher initial capital cost can sometimes be the economically superior option. Automation is not merely changing the cost of moving a case. It can change the amount of productive capacity available per square foot.

Payback Period Can Hide as Much as It Reveals

Warehouse executives understandably want a payback number. Industry expectations can be demanding. An MHI discussion of robotics ROI noted that companies commonly look for a return in roughly two years, although appropriate targets vary considerably by operation and technology.

There are deployments demonstrating that kind of return. Davcor Group, for example, reports achieving ROI on its AutoStore implementation in under two years. Its operation handles approximately 400 orders daily, with the system performing roughly 1,200 picks and 200 to 300 put-aways per day and offering capacity of up to 600 picks per hour. But payback should not become the only lens.

Suppose Solution A has a two-year payback but will struggle with projected SKU growth in year four. Solution B takes three years to pay back but provides modular expansion, substantially higher storage density, and greater peak capacity. Which one has the better economics? The spreadsheet may initially favor Solution A. The operation may eventually favor Solution B.

Automation Has Hidden Costs Too

The same logic has to work in the opposite direction. Automation projects can underperform financially when companies underestimate integration complexity, implementation disruption, or ongoing operating costs. An automated system performing well in isolation does not guarantee a high-performing warehouse. Robots still need orders. AS/RS equipment needs inventory information. Picking stations need balanced workflows. WMS, warehouse execution systems, controls, robotics, conveyors, and humans all have to operate as an integrated system.

Downtime also becomes economically important. As a facility becomes more automated, failures can affect larger portions of throughput. Maintenance capabilities, spare-parts strategies, system redundancy, vendor support, and recovery procedures should therefore be considered during financial justification, not after commissioning. This is where some ROI models become overly optimistic: they model the technology but not the operating environment around it.

A Better Way to Calculate Automation ROI

Warehouse operators should evaluate automation across at least five economic dimensions:

Labor: How many labor hours disappear, and how many are simply moved elsewhere?

Throughput: How much additional volume can the existing operation process?

Space: Can higher storage density postpone an expansion or relocation?

Service: What is the financial effect of improved accuracy, shorter cycle times, or later order cutoffs?

Risk and flexibility: Can capacity scale up or down as volumes, SKUs, and customer requirements change?

These factors create a more realistic picture of value than dividing an equipment purchase by annual labor savings.

The Bigger Question

The economics of warehouse automation have matured along with the technology. Today, the strongest business cases are not necessarily those promising the largest reduction in headcount. They are the ones connecting investment to a broader operating model: labor productivity, throughput, capacity, space utilization, service performance, scalability, and risk. For warehouse leaders, that requires a shift in thinking. Do not ask only how fast automation pays for itself. Ask what the warehouse will be capable of once it does. That is where the real economics of warehouse automation begin.

The post Beyond the ROI Spreadsheet: The Real Economics of Warehouse Automation appeared first on Logistics Viewpoints.

Google has identified a problem with artificial intelligence agents that deserves the attention of supply chain and logistics executives. Increasingly capable AI systems can become distracted by outdated information, pursue work nobody requested, and consume substantially more resources without advancing their assigned objectives.

In research published November 6, 2026, Google senior staff software engineer Sam Saccone and senior staff data scientist Madhuri Garikapati examined coding-agent interactions across thousands of engineers and hundreds of millions of individual agent steps.

Their findings were striking. After controlling for factors including developer, task type, model tier, and codebase complexity, sessions involving unprompted detours into unrelated software bugs were associated with 70% more agent steps, 77% more tokens consumed, and 105% more waiting time.

Google calls this phenomenon agent distraction.

Although the research examined software-development agents, I believe it exposes a broader issue that supply chain technology providers and end users need to consider as autonomous systems move into operational environments.

An AI agent can be technically productive while being operationally counterproductive.

That distinction has significant implications for the economics of agentic AI. Every unnecessary investigation consumes computing resources and execution time. More importantly, it can delay the operational improvements that justified the investment in the first place.

How AI Agents Lose Their Way

Google’s researchers focused on a seemingly minor source of technical debt: old TODO and FIXME comments embedded in software code.

Developers leave these comments to identify problems or unfinished work. Over time, some become obsolete. The underlying bug may have been corrected, the requirement abandoned, or the original issue superseded by another solution.

Human developers can recognize many of these references as historical information. AI coding agents sometimes treat them as reasons to initiate additional work.

Google found that approximately 13% of examined file-reading sessions exposed agents to these bug-related references. In 6% of those encounters, agents subsequently acted on the referenced bug. Of those actions, 83% were not requested by the developer.

Among the unprompted detours, 41% concerned already-closed tickets. Another 57% involved bugs that remained open but fell outside the developer’s assignment.

The issue is not simply wasted computing capacity. Agents can investigate irrelevant problems, make unnecessary changes, and introduce additional complexity while the original assignment remains unfinished.

Google responded by developing specialized agents to identify and remove obsolete software references. The effort eliminated thousands of outdated references over six months, with more than 95% acceptance from human reviewers.

That response is important. Google did not conclude that technical maintenance should be abandoned. It demonstrated that AI can perform maintenance effectively when that work is itself the authorized assignment.

The distinction between legitimate improvement and unrequested activity is central to the challenge.

When Technical Debt Becomes Operational Debt

Supply chain environments contain their own versions of obsolete software comments.

Transportation management systems accumulate historical carrier configurations, routing rules, and exception records. Warehouse management systems retain legacy procedures and inventory adjustments. Procurement applications contain supplier qualifications, contract histories, and outdated operating instructions.

These have traditionally been data quality and systems maintenance problems. Agentic AI introduces another dimension.

When autonomous agents interpret outdated operational context as instructions to initiate work, technical debt can become operational debt.

Consider an AI transportation agent assigned to resolve shipments at risk of missing delivery commitments.

The agent identifies available capacity, evaluates carrier alternatives, and calculates recovery options. During that process, it encounters a historical note about a carrier integration problem.

The issue may have been resolved years earlier. It may have no relevance to the affected shipments.

Nevertheless, the agent could begin investigating historical transactions, running diagnostics, or proposing integration changes while the shipments remain unresolved.

This is an illustrative scenario, not a result measured in Google’s study. But the underlying risk is credible wherever autonomous systems rely on large collections of operational information.

The challenge becomes more consequential when multiple agents interact. A transportation agent pursuing an irrelevant issue might involve inventory, procurement, or customer service agents, spreading unnecessary work across the organization.

The question for supply chain executives is therefore not simply whether agents can reason, communicate, and execute tasks. It is whether they can consistently distinguish between the work they were assigned and other work they happen to discover.

Five Rules for Keeping AI Focused on Results

Traditional supply chain management already provides much of the operating discipline needed to address this challenge. Organizations establish objectives, approve plans, manage exceptions, control spending, and measure results.

Those principles should govern autonomous agents just as they govern human operations.

1. The Current Mission Takes Precedence

Every agent needs a clearly defined operational objective.

If an agent is assigned to resolve 200 delayed orders, success means moving those orders toward resolution within established service, cost, and quality constraints.

The agent may identify unrelated process deficiencies. Those findings can be documented without interrupting the current assignment.

This distinction becomes critical as agents gain access to software-development tools, enterprise applications, and other agents.

The operating objective, not the expanding universe of possible improvements, must determine priorities.

2. Frozen Decisions Must Remain Frozen

Supply chains operate through approved plans, sourcing strategies, replenishment parameters, and financial controls.

An agent should not independently reopen those decisions simply because another approach appears technically attractive.

An inventory agent operating under approved replenishment parameters might identify an alternative optimization methodology. That observation does not authorize it to redesign the replenishment strategy during execution.

New evidence of material operational, safety, compliance, or financial risk may justify reconsideration. Otherwise, approved decisions should remain authoritative until the appropriate owner changes them.

Autonomy to execute is not authority to redefine the business.

3. Diagnose Demonstrated Blockers, Not Speculative Problems

Enterprise systems are imperfect, and AI agents will inevitably encounter inconsistent records, historical exceptions, and technical deficiencies.

Not every imperfection requires immediate investigation.

A warehouse interface failure preventing shipment confirmation is a genuine blocker. An unrelated historical data inconsistency may warrant maintenance, but not necessarily during an active fulfillment operation.

Agents should distinguish between issues preventing execution and those better addressed separately. Credible safety, security, and compliance risks still require appropriate escalation.

The existence of a problem does not automatically make solving it part of the current assignment.

4. One Blocker, One Bounded Investigation

AI agents can be persistent problem solvers. Unfortunately, persistence can become an endless engineering exercise.

An agent investigates an error, discovers another issue, proposes a repair, and initiates another investigation. Eventually, the original objective disappears behind a growing chain of technical activity.

I think of this as an AI engineering doom loop.

The remedy is to establish a defined investigation scope, resource limits, and stopping conditions.

If the blocker is resolved, execution resumes. If it remains unresolved, the agent reports the evidence and escalates rather than independently expanding the investigation.

A transportation exception should not automatically become a TMS modernization project.

5. Measure Operational Output, Not Agent Activity

This is the most important principle.

An AI system can execute thousands of steps, generate hundreds of analyses, and consume substantial computing resources without delivering a meaningful business result.

For transportation, the relevant outcomes include on-time delivery, cost per shipment, and exception-resolution time. For warehousing, they include throughput, order accuracy, and labor productivity. For procurement, they include completed sourcing events, verified savings, and reduced supply exposure.

These measures must account for quality, safety, and financial performance.

The number of tasks an agent completes is not the same as the amount of value it creates.

What This Means for Supply Chain Technology Investments

Google’s research should influence how enterprises evaluate agentic AI solutions.

Technology suppliers increasingly emphasize autonomous execution, reasoning capabilities, and the number of workflows their agents can perform. Those capabilities matter, but they do not establish economic value by themselves.

A transportation agent that reduces manual planning effort but repeatedly diverts into unrelated investigations may deliver less value than a simpler system that reliably completes its assigned work.

The same applies to warehouse exception management, inventory reconciliation, and procurement automation.

For technology suppliers, agent reliability and disciplined task completion should become important areas of differentiation. For end users, pilot evaluations should measure completed business transactions, unnecessary tool activity, scope deviations, human interventions, and elapsed time to resolution.

Organizations should also distinguish between the cost of model inference and the total cost of achieving a business result.

A lower-cost model that repeatedly pursues irrelevant work may ultimately be more expensive than a higher-cost model that completes the assignment efficiently.

The relevant economic measure is cost per successfully completed business outcome, not simply cost per token or agent execution.

That requires governance extending across the entire agent workflow, including access to authoritative information, explicit operating boundaries, financial limits, human approval for consequential decisions, and audit trails connecting actions to outcomes.

These controls are not obstacles to autonomy. They are what make autonomy commercially useful.

The Bottom Line

Supply chain organizations have spent decades investing in technology to improve the movement of materials, products, information, and capital. Agentic AI offers another opportunity to improve those operations by coordinating decisions and executing work that previously required substantial human intervention.

But Google’s findings suggest that greater agent capability does not automatically produce greater operational discipline. As AI systems become more autonomous, their ability to pursue irrelevant work may grow alongside their ability to perform valuable work.

The successful agentic AI platforms will not necessarily be those with the most sophisticated reasoning or the greatest number of autonomous capabilities. They will be those that consistently complete the right work, respect established operating constraints, and deliver measurable economic value.

The objective is not to build perpetual AI projects. It is to build better supply chain operations.

Research source: Sam Saccone and Madhuri Garikapati, Google Developers Blog, Agent Distraction, November 6, 2026.

The post When AI Agents Lose Focus: What Google’s Research Means for Supply Chain Operations appeared first on Logistics Viewpoints.