Supply Chain Analytics Pulse

A Fleet Executive’s Playbook for the Next Fuel Shock 

July 2026 | Supply Chain Analytics Pulse

In late February 2026, the Strait of Hormuz, which carries roughly a fifth of the world’s seaborne oil, effectively shut down as a sudden regional escalation swept the Gulf. Brent crude spiked from around 69 dollars a barrel to nearly 120 within weeks, with traders briefly pricing in scenarios above 150. The International Energy Agency called it one of the largest single supply shocks the oil market has ever recorded.

It was also, in a sense, not the important story. The Red Sea has been under similar strain since November 2023, when Houthi attacks forced most container shipping onto a longer route around the Cape of Good Hope, a detour still in effect today. Carriers testing a return to Suez this spring abandoned those plans within days once the Iran war escalated, two unrelated crises compounding each other. Go back further and the pattern holds: a grounded ship shut the Suez Canal for six days in 2021, and drought cut Panama Canal transits through 2023 and 2024. War, militants, an accident, a drought, different causes, the same signature: a shock that holds for months or years, then eases just long enough for the next one to arrive from a different direction. 

01 Managing Fuel Risk on Purpose

If the trigger cannot be predicted, the only thing left to manage deliberately is exposure. That is a solved problem in other fuel-intensive industries, and most of the solution is not new technology. It is underused financial discipline.

Large fuel consumers, airlines, shipping lines, and some trucking fleets, typically draw on three instruments.

• Futures and swaps lock in a fixed price for future consumption: if prices rise, you win, and if they fall, you are stuck paying above market, which is the real cost of certainty.
• Call options work more like insurance: you pay a premium for the right, not the obligation, to buy at a set price, so if prices spike the option pays off, and if they fall instead you lose only the premium and still benefit from the cheaper market price.
• Surcharge and fuel-recovery programs are a third path entirely, a pass-through mechanism tied to real, current pricing rather than a fixed hedge position, though several fuel-risk consultancies note that a surcharge program can quietly stop matching real exposure if it is not re-indexed against current benchmarks.

02 Electrification & Wider Menu Options

Electric trucks have moved from a sustainability talking point to something industry leadership is willing to say in financial terms. Geotab’s CEO argued that resilience s a third pillar of the electric vehicle business case, alongside emissions and total cost of ownership. It is a real, attributable argument, and it sits next to real constraints the IEA’s Electric trucks still cost roughly two to three times as much to purchase as diesel equivalents, and battery weight creates genuine payload limitations for high-weight freight, sometimes forcing additional vehicle movements that partly offset the fuel savings.

Where the economy already works cleanly is return-to-base, regional, and last-mile routes with predictable duty cycles. Workhorse’s step-van operations reportedly saved about 42.5 cents per mile against gasoline equivalents, using electricity priced near 11 cents per kilowatt-hour against gasoline near 2.98 dollars a gallon, savings that get larger, not smaller, exactly when a fuel shock hits.

Long-haul remains the harder case, where energy density and existing diesel infrastructure still dominate for now.

But electrification is one option on the menu, not the only item on it.

• Renewable diesel is a drop-in substitute for petroleum diesel that needs no new trucks or fueling infrastructure at all, only a different feedstock at the pump, and it has served several fleets as bridge fuel while electric infrastructure catches up.
• Compressed and renewable natural gas is a third pathway that some of the largest fleets in the country have bet on directly.
• Hydrogen fuel cells remain earlier stage for heavy trucking but are drawing serious investment for exactly the same reason electrification is: they remove a fleet from a single fuel market’s price swings.

The strategic point is not which of these wins. It is that a fleet running on a single fuel source has no internal fallback when that fuel spikes, while a fleet with two or three live pathways can shift volume toward whichever one is cheapest or most available in a given quarter. Choosing which pathway to expand, and when, is itself a decision made under uncertainty, which is a problem finance already has tools for.

03 Why Waiting isn’t Free Either

The standard finance tool for a large, irreversible capital decision is net present value: estimate the cash flows, discount them, invest if the number comes out positive. NPV implicitly assumes an invest-now-or-never choice and ignores the value of waiting for uncertainty to resolve. Real options analysis is the academic framework built to close that gap. It treats the ability to delay, expand, or abandon an investment as having quantifiable value, the same way a financial option does, and it is already an active research area applied specifically to fleets: a 2023 review of real options in transportation research describes work modeling fleet investment decisions under regulatory uncertainty, and references earlier research building a Monte Carlo model to optimize fleet replacement timing specifically under fuel price uncertainty.

Classic real options logic says that higher volatility increases the value of waiting. You would rather defer an irreversible bet until the picture clears, which is the entirely reasonable instinct behind fleets that held off on electric purchases through 2026’s turmoil. But that logic only holds if the investment itself does not change your exposure to the uncertain variable. Fleet electrification is unusual in exactly this respect: it does not bet on where oil prices go, it exits that market almost entirely. That reframes the real question a finance team should be modeling. Not “will fuel be more expensive next year,” but “what is removing fuel-price variance from our cost base worth, independent of which direction prices move”. Those are different calculations, and only the second one avoids requiring anybody to correctly predict Middle East geopolitics.

Running that second calculation well, at the level of individual routes and duty cycles rather than fleet-wide averages, used to require a dedicated quantitative team. That constraint is loosening, which is the subject of the next section.

04 Forecasting Risk the Way We Already Forecast Demand

The same predictive capability that has reshaped demand forecasting is being pointed at fuel and disruption risk, and the mechanics transfer more directly than most executives assume. In aviation, where fuel is typically the largest single cost line, procurement teams are already using forecasting layers that ingest historical price data from indices like Platts or OPIS alongside exchange rates, planned schedules, and weather signals, producing not just a price forecast but a “what if” simulator that lets a buyer test purchasing windows against different scenarios before committing. There is nothing airline-specific about that architecture. A diesel-buying trucking fleet has the same categories of input available, historical price data, route and utilization schedules, weather, and increasingly, real-time shipping and geopolitical signals.

Monte Carlo simulation of investment timing used to be the domain of a company with its own quantitative finance team. AI-assisted modeling tools are making a simplified version of that same analysis reachable for a mid-size fleet’s own finance function, which matters because the harder part of real options thinking was never the concept, it was the computation. The second is earlier warning. Natural language processing applied to shipping data, vessel tracking, and war-risk insurance pricing can flag a building disruption in a shipping lane weeks before it shows up as a line item on a fuel bill, turning the governance triggers from a quarterly review into something closer to a live dashboard.

05 Case Studies From the Field

06 Resilience Readiness Check: What to Ask Yourself Now

A short set of questions worth answering honestly before the next disruption, whatever form it takes, arrives.

07 Five Steps to Get There: Choose Your Perspective

Click on the perspective you want to analyze – executives or analysts.

06 Conclusion

Executives do not get graded on whether they called the next disruption correctly. Nobody called two and a half years of Houthi missiles reshaping container shipping, and nobody called a handful of carriers abandoning their return to Suez within days because a different war, in a different strait, escalated at the wrong moment. The specific chokepoint always changes. What does not change is that somewhere, a narrow passage the world had started to treat as permanent will fail again, on a timeline nobody can give you.

The fleets that come out ahead of the next one will not be the ones with the best forecast. They will be the ones for whom the forecast matters less, because a hedge book was sized to real risk tolerance instead of habit, a fuel mix did not collapse the moment one pathway got expensive overnight, and a monitoring system flagged trouble in a shipping lane before it ever showed up on a fuel invoice. That is a duller kind of advantage than being right about geopolitics. It is also the only kind that keeps working no matter which conflict, drought, or accident causes the next one.

Supply Chain Analytics Pulse

What Is Hyperautomation, and Why It’s Becoming Supply Chain’s Operating System

July 2026 | Supply Chain Analytics Pulse

By 2034, the global hyperautomation market is projected to reach $235 billion, up from $55.54 billion in 2025 – a 17.38% CAGR that outpaces almost every other enterprise software category. But the number that should actually worry supply chain leaders isn’t the market size. It’s this one: Gartner expects that in 2026, 30% of enterprises will automate more than half of their network activities, up from under 10% in mid-2023. That’s a tripling of automated network coverage in three years, and most supply chain organizations are not on that curve.

Hyperautomation is not RPA with a new label. It’s the orchestration layer that turns isolated bots, machine learning models, and point solutions into a single, self-correcting operating system for the business. For supply chains specifically, that shift is arriving at the same moment as agentic AI, and the two are converging into something the research increasingly treats as inevitable rather than optional.

01 What Hyperautomation Actually Is — and Isn’t

Gartner coined the term several years ago, defining it as “a business-driven, disciplined approach that organizations use to rapidly identify, vet, and automate as many business and IT processes as possible.” The definition has aged well, but the technology stack behind it has changed completely.

INFORMS’ Analytics Magazine traced the lineage back to 2020–2021: RPA automated individual, rule-based tasks. Hyperautomation was framed at the time as “the extension of legacy business process automation beyond the confines of individual processes”, essentially automating the automation itself, by dynamically discovering business processes and then generating the bots needed to run them. That framing was directionally right but technologically premature. RPA in 2020 still broke every time a button moved on a screen.

What’s changed is the cognitive layer sitting on top of the mechanical one. As KPMG Belgium puts it in its 2026 analysis, hyperautomation today orchestrates “RPA, AI, ML, process mining, intelligent document processing, and more” to automate entire end-to-end business processes — not tasks, not even processes in isolation, but the connective tissue between them. KPMG’s central argument is worth sitting with: hyperautomation isn’t the destination. It’s the foundation that makes agentic AI possible, because you cannot hand autonomous decision-making to an AI agent operating inside a process nobody has mapped, running on data nobody has cleaned, in a system nobody has connected to anything else.

RSM’s 2025 analysis highlights what’s newly possible: agentic UI automation, or “computer use.” Unlike scripted bots that click fixed screen coordinates, vision-plus-LLM agents interpret a live screen, decide the next best action, and adapt when interfaces change. This expands automation beyond API-enabled systems to nearly any software a human can use, unlocking significant value for supply chain organizations that still rely on legacy TMS, WMS, and supplier portals not designed for modern integrations.

The market data underscores the urgency. Fortune Business Insights values the global hyperautomation market at $55.54 billion in 2025 and projects it will reach $235 billion by 2034. Robotic process automation remains the largest segment (25%), while machine learning (20%) is growing fastest by making automation smarter, not just faster. Manufacturing accounts for only about 8% of hyperautomation spending, trailing sectors like BFSI (15%) and IT & Telecom (14%) despite having significant potential to benefit.

02 What the Research Is Finding

Four recent studies — spanning incident management, big data integration, procurement, and AI ethics — converge on a consistent pattern: the technical capability for end-to-end automated supply chain decision-making now exists, but the organizational and governance infrastructure to use it responsibly is lagging behind.

Incidents are moving from reactive to causal.

Jingar’s 2023 framework for agentic AI-based incident intelligence showed that specialized AI agents using collaborative reasoning over a causal knowledge graph reduced mean time to root-cause identification by 30% and improved incident resolution accuracy by 22% compared with traditional methods. More importantly, the agents uncovered hidden dependencies across suppliers, transportation, and operations that rule-based systems missed.

The AI/ML/deep learning stack is becoming standard infrastructure, not a differentiator.

Husnain’s 2026 review argues that big data, AI, and deep learning work together as one decision-making system: big data provides the inputs, AI finds patterns and optimizes, and deep learning handles unstructured data like images, sensor feeds, and text. The key takeaway is matching AI to the decision horizon – AI-driven scenarios for strategic choices, predictive analytics for tactical planning, and real-time AI for operational decisions such as rerouting and warehouse activity.

Procurement is quietly becoming the most automatable function in the supply chain.

Ilamurugan’s 2026 review outlines a five-stage ML-driven spend analysis workflow: data ingestion, cleansing, classification, pattern recognition, and predictive analytics, and reports procurement cycle reductions of over 30% in pharmaceutical supply chains. It also points to near-term autonomous procurement agents that can negotiate with vendors, place orders, and manage compliance using reinforcement learning, supported by technologies like NLP and blockchain.

None of this is ethically neutral, and treating it as such is where the risk concentrates.

Raikar et al.’s 2026 study provides a needed counterbalance, showing that while AI improved order fulfillment by 20–50% and reduced logistics costs by 15–20%, it also introduced service-level disparities due to biased historical training data. The authors recommend embedding fairness constraints directly into optimization models, reducing discriminatory outcomes from 20–40% to 5–10% while sacrificing less than 3% of revenue gains.

Taken together, the four papers describe the same arc from three different angles: automation is moving from detecting problems to reasoning about their causes, from single-function tools to integrated decision infrastructure, and from human-supervised execution to governed autonomy, with the ethical and governance layer as the current bottleneck, not the technology itself.

03 Four Companies Already Living This

04 Hyperautomation Maturity Assessment

Before benchmarking against any of the case studies above, supply chain leaders need an honest answer to a harder question: where does the organization actually sit today? Drawing on Gartner’s five-level maturity framework (Awareness → Active → Operational → Systematic → Transformational) and cross-referencing it against sector-specific evidence from logistics maturity research, the table below gives a working self-assessment across five dimensions that determine whether hyperautomation investment will compound or stall.

Most manufacturing and logistics organizations in 2025–2026 sit at Level 2 (Active) on technology adoption and Level 1 (Ad Hoc) on governance – and it’s the governance gap, not the technology gap, that most consistently derails hyperautomation programs, echoing Raikar et al.’s finding that fairness and accountability infrastructure lags behind technical capability by a wide margin.

A simple self-check: if routing compliance reports take over a week, AI pilots haven’t reached production, or no one owns fairness and bias monitoring, the organization is likely at Level 1 or 2, regardless of its automation investments. As maturity research shows, owning automation software is not the same as achieving operational maturity.

05 Four Steps to Get There: Choose Your Perspective

The path to hyperautomation is a sequence, and skipping steps produces exactly the kind of “sunk cost with no ROI” outcome the market research keeps flagging as the leading cause of stalled automation programs. Below are four steps tailored to two different roles in the organization.

Click on the perspective you want to analyze – executives or analysts.

06 Conclusion

The technology described above is no longer speculative. Agentic AI is identifying root causes across supply chain layers that rule-based systems physically cannot see. Machine learning is compressing procurement cycles by a third. Digital twins are cutting quality defects by 30% over four years at named factories, not hypothetical ones. Risk-sensing models are turning a one-day response plan into a two-hour one, with a documented save during an actual geopolitical shock.

What’s still catching up is everything sitting underneath the technology: the process discipline to know what’s actually being automated, the data infrastructure to trust the outputs, and – the piece the research is most consistent about – the governance architecture to ensure that efficiency gains aren’t quietly extracted from the communities and customer segments least equipped to notice. Gartner’s own maturity data suggests most organizations are one or two levels below where their technology purchases imply they should be. Closing that gap, deliberately and in sequence, is what turns hyperautomation from a line item into the operating system the market size numbers assume it will become.

Future of Warehousing: RFID, IoT, and Vision Picking

July 2026 | Supply Chain Analytics Pulse

The warehousing industry is at the forefront of a technological transformation. Technologies such as radio-frequency identification (RFID), the Internet of Things (IoT), and augmented reality (AR) are reshaping warehouse operations and redefining the standards of efficiency, precision, and inventory visibility. These technologies are no longer simply operational tools—they are becoming strategic assets that enable warehouses to meet the growing demands of global supply chains.

RFID systems, combined with IoT capabilities, enable seamless inventory tracking, reducing human error and optimizing workflows. At the same time, augmented reality technologies, such as smart glasses, provide warehouse employees with real-time information and hands-free support, significantly improving order-picking efficiency. From autonomous drones equipped with RFID readers to vision-picking systems that minimize errors, these innovations are transforming warehouses into smarter and more resilient environments.

01 RFID and IoT: Building the foundation for warehouse visibility

Unlike traditional barcode systems, RFID technology enables remote and simultaneous scanning without requiring a direct line of sight. This eliminates the need for manual close-range scanning and allows warehouses to track inventory more efficiently. RFID tags function as digital identifiers attached to individual products or entire pallets, storing information that is continuously updated as goods move through different warehouse processes.

How RFID tracking works

• Shipments arrive at the warehouse.
• RFID tags are attached to products or pallets.
• Information is stored and continuously updated.
• Data is transmitted to the warehouse management system.
• Inventory is tracked in real time.

Connected devices, sensors, and warehouse management systems extend this visibility beyond inventory tracking, enabling asset management, warehouse automation, and real-time monitoring. These technologies have become increasingly important for industries that rely on fast and accurate order fulfillment, including e-commerce, logistics, and manufacturing.

As supply chains continue to grow in complexity, warehouses increasingly depend on real-time information to maintain efficiency and adapt to changing market requirements.

02 From warehouse shelves to autonomous drones

One example of this transformation is the partnership between Verity, On, and Maersk, which explored how RFID technology and autonomous inventory tracking could improve warehouse visibility and operational efficiency. The pilot project combined RFID capabilities with Verity’s AI-driven drones to bridge the gap between physical warehouse operations and digital supply chain management.

The drones navigated warehouse aisles autonomously, scanning RFID tags and identifying products throughout different stages of the warehouse lifecycle. Unlike traditional RFID systems, which depend on fixed checkpoints and close-range scanning, the drones enabled inventory tracking beyond the line of sight and allowed the precise identification of items inside and outside boxes.

These results demonstrate how RFID-enabled automation can enhance inventory visibility, transparency, and accuracy without disrupting existing warehouse operations.

03 What research tells us about RFID

Scientific research further highlights the role of RFID in warehouse and logistics management. According to the study Role of Radio Frequency Identification (RFID) in Warehouse and Logistic Management System using Machine Learning Algorithm, RFID technology contributes to the analysis of warehouse data and the generation of actionable insights based on historical trends and future forecasts. Depending on organizational resources and capabilities, the study reports accuracy levels of up to 99.57%.

Research highlights three key contributions of RFID systems:

Automated inventory handling and retrieval

Additional studies demonstrate how RFID readers positioned at warehouse entrances and exits can track incoming and outgoing goods, while warehouse management systems coordinate inventory movements automatically. Simulations conducted as part of the research confirmed the efficiency of these systems in locating and retrieving products.

04 Vision picking and the role of smart glasses

Automation continues to reshape warehouse operations, but order picking remains one of the activities that still depends heavily on human flexibility. Augmented reality technologies, particularly smart glasses, aim to support workers by reducing time-consuming secondary tasks and providing real-time information directly within their field of view.

Smart glasses integrate text, graphics, and video into the physical environment, allowing employees to receive instructions, view picking statuses, and navigate warehouses without interrupting their workflow. Unlike virtual reality systems, augmented reality solutions maintain a continuous connection between workers and their surroundings.

Why vision picking matters

• Real-time instructions
• Hands-free operations
• Reduced search times
• Fewer picking errors
• Better ergonomic support

05 Impact on productivity

Several case studies illustrate the impact of smart-glass technologies on warehouse performance:

A case study conducted in two warehouses in Belgium also demonstrated that successful implementation depends not only on technology itself but also on employee involvement. The warehouse that actively engaged employees in improving the system achieved significantly better results.

06 Looking ahead

The integration of RFID, IoT, and augmented reality represents more than a technological upgrade for warehouses. These technologies are transforming warehouses from static storage facilities into dynamic, data-driven ecosystems capable of predicting, adapting, and responding to market demands in real time.

Organizations that adopt these technologies are not only optimizing warehouse operations but also redefining their supply chain strategies and creating new opportunities for inventory planning and logistics management.

The future of warehousing lies in the convergence of technology and strategy, where visibility, adaptability, and operational intelligence become essential competitive advantages.

According to market experts, only 22% of companies measure savings after spending actually occurs. In an industry where margins run at 1–2% for FTL operators and groupage networks routinely bleed on lanes nobody is actively monitoring, that measurement gap alone represents a significant portion of unrealized value.

Understanding exactly where the gap opens, how it compounds, and what the operational infrastructure required to close it actually looks like is more useful than any individual cost-reduction initiative.

In case you want to jump ahead

1. The Architecture of the Saving Gap

2. Maersk: The Custom Control Tower & the Execution Distinction

3. Additional Industry Cases

4. Organisational Readiness Check

5. Key Implementation Steps: Choose Your Perspective

6. Conclusion

The Architecture of the Saving Gap

The savings gap is not a single failure point. It is the compound effect of three structural mechanisms that operate simultaneously in almost every logistics organization and understanding them precisely matters, because each one requires a different intervention to close.

Monitoring Towers VS Execution Towers

The dominant mental model of a supply chain control tower is a visibility tool, a centralized dashboard aggregating shipment data, flagging exceptions, and surfacing performance metrics across lanes and carriers. This architecture is genuinely useful, but it also does not close the savings gap.

A monitoring control tower tells a dispatcher that a shipment is running at 61% truck utilization and that a consolidation-eligible load exists on the same lane with 18 hours of remaining SLA buffer. An execution-connected control tower holds the first shipment, triggers the consolidation, and routes both under the preferred carrier rate automatically, within defined rules, without requiring a human decision at the transaction level.

The practical implication is that two organizations can purchase identical control tower software and achieve fundamentally different margin outcomes, because one has connected the tower’s output to automated routing logic and the other has connected it to a dashboard that someone checks in the morning.

Organizations using AI-enabled supply chain control towers have reported up to 30% reduction in inventory and 8% lower logistics costs.

Static Planning VS Continuous Planning

Most logistics cost-saving analysis is conducted on a periodic basis, which creates an inherent structural problem: the findings reflect a snapshot of a market that continues to move. A consolidation opportunity identified in a September lane review may not exist in the same form in January when the planning assumptions have shifted.

Cost management is the most critical priority for one-third of corporate leaders globally in 2025, according to BCG, but the majority of logistics operators are still managing cost through planning cycles that update annually or quarterly rather than continuously. The result is that identified savings have a decay curve: a saving found in Q1 is partially realizable by Q3 and may be substantially eroded by Q4, precisely when the implementation that was meant to capture it completes. 

The technical shift required is from planning as a periodic project to planning as a continuous operational function, where optimization algorithms run against live TMS and ERP data, route logic updates when carrier availability changes, and consolidation rules adjust dynamically when SLA buffer patterns shift. 

The Compliance Enforcement Gap

Even when a saving is correctly identified, implementation timelines are respected, and routing logic is updated in the TMS, the final failure point is compliance enforcement: the systematic, measurable assurance that eligible shipments are actually moving under the conditions the analysis prescribed.

A contracted carrier rate is a potential saving that becomes actual only when that carrier is consistently selected for eligible shipments rather than spot alternatives. In most TMS environments, the routing guide exists, the preferred carrier is defined and a meaningful percentage of eligible volume moves outside it for reasons that are never tracked. Those reasons range from carrier capacity unavailability to dispatcher preference to system configuration errors. All of them represent the gap between the rate the procurement team negotiated and the cost that appears on the freight invoice.

According to Deloitte research, 76% of manufacturers report limited visibility across their supply chains and that is precisely the condition that allows compliance failures to compound below the threshold of detection. The compliance enforcement gap is the last and most operationally tractable expression of the savings gap.It requires measurement infrastructure that almost no organization has built.

Maersk: The Custom Control Tower & The Execution Distinction

The Maersk Customs Control Tower case is the clearest documented illustration available of the difference between a monitoring architecture and an execution architecture and of what that difference produces in financial terms.

One of Maersk’s global clients had many stakeholders running different customs processes, with low visibility of customs data and limited ability to understand the business implications and make the right corporate decisions. This is an exact description of all three mechanisms operating simultaneously: 

The architecture Maersk implemented centralises customs operations with automatic validation of data and documents to identify discrepancies before distributing orders to brokers, identifies and resolves exceptions automatically based on pre-defined business rules, and logs all exceptions continuously for further analysis and ongoing process improvement. 

The operational significance of that architecture is in what it removes from the human decision loop. Document validation, exception identification, broker routing, and duty classification consistency are not decisions that a dispatcher or customs coordinator reviews case by case, but enforced by rules that apply to every declaration, automatically. The compliance enforcement gap closes, because the system architecture does not permit non-compliance on the decisions it governs.

Visibility without enforcement produces insight. Enforcement without visibility produces rigidity. The combination produces realized savings.

With a $250,000 investment in the Customs Control Tower, the client achieved $1.5 million in reduced annual spend on customs services and $15 million in reduced duties paid in the first 12 months

Customs is a specific process domain, but the mechanism it illustrates is universal. The same logic applies to carrier selection in FTL networks, consolidation decisions in groupage operations, and inbound call-off timing in manufacturing supply chains. In each case, the saving is present in the data before any new analysis is commissioned. What is typically absent is the enforcement architecture that captures it.

Additional Industry Cases

The execution distinction illustrated by Maersk is not specific to customs or to shipping. The same structural gap and the same architectural response appears across industrial manufacturing and road freight operations, each with its own operational expression of the same underlying problem.

Click on the use case you want to learn more about:

Before committing to control tower investment, planning automation, or optimization tooling, supply chain leaders should assess their current architecture across four dimensions. Most LSPs and manufacturers in 2025–2026 sit at the Developing stage on data infrastructure and Early Stage on compliance enforcement and it is the enforcement gap that is hardest and most consequential to close.

The developments described above are not sequential innovations. They are converging simultaneously.

Click on the perspective you want to analyze – executives or analysts.

Conclusion: The Infrastructure of Realisation

Traditional cost programmes were designed to identify savings. The organisations closing the savings gap are designed to capture them. The difference is whether the conclusions of analysis are encoded into the operational system that processes every shipment, automatically, with audit trails that prove the encoding is holding.

The Maersk case makes this concrete at the LSP level: $16 million in annual return from a $250,000 investment, not because new savings were discovered, but because existing ones were finally built into the architecture processing every customs transaction. The BMW case makes it concrete at the manufacturing level: premium freight costs reduced not by renegotiating carrier rates, but by making supply risk visible early enough to respond through planning rather than emergency logistics.

For a structured view of where the specific cost leakages most commonly occur across Groupage, FTL & 4PL  explore the Log-hub Cost Savings Initiative.

At Renault, a single daily planning cycle involves 45,000 decision variables and 100,000 constraints, and the system is expected to produce a high-quality answer in under five minutes. Behind it sits an optimization solver: a mathematical engine that evaluates trade-offs at a scale no planning team could match manually.

In case you want to jump ahead

1. Inside the Solver: Different Paths to the Same Goal 

2. The Rise of Hybrid Optimization

3. Behind the Solver: Renault’s Packaging Management System

4. Evaluating Optimization Through Different Lense

5. Supply Chain Optimization Maturity Self-Assessment

6. In Conclusion

SUPPLY CHAIN EXAMPLE

SUPPLY CHAIN EXAMPLE

SUPPLY CHAIN EXAMPLE

SUPPLY CHAIN EXAMPLE

SUPPLY CHAIN EXAMPLE

The Rise of Hybrid Optimization

Instead of relying exclusively on linear programming, heuristics, or constraint programming, modern platforms orchestrate several approaches simultaneously, allowing each method to contribute where it performs best. 

Depending on the problem, they may use linear optimization to evaluate strategic trade-offs, combinatorial optimization to manage complex allocation decisions, and heuristics to accelerate the search for high-quality solutions. Rather than relying on a single algorithm, solvers can switch between or combine approaches to balance solution quality, computation time, and model complexity. 

Behind the Solver: Renault’s Packaging Management System

At Renault Group, reusable packaging continuously circulates between approximately 1,400 suppliers, 40 plants and cross-docks, and multiple cleaning, repair, and recovery locations. Ensuring that the right packaging is available at the right place and time is essential for maintaining production continuity while controlling transportation and asset-related costs.

For each day, supplier, plant, and packaging type, Renault must determine how many packaging units should be moved throughout the network while respecting inventory balances, shipment requirements, facility capacities, and operational constraints. At the same time, the system must balance several potentially conflicting objectives, including minimizing packaging shortages, reducing the number of shipments, and limiting overall travel distances.

To support these decisions, Renault developed a Packaging Management System (PMS) powered by Hexaly.

A typical optimization model includes approximately:

<1.5%

avarage optimal gap

The challenge is finding a high-quality solution within a timeframe that allows planners to act on the results.

According to Hexaly, the system achieves an average optimality gap below 1.5% within a five-minute solving window, enabling Renault to regularly optimize packaging flows at a scale that would be impractical through manual planning or spreadsheet-based analysis.

Optimization solvers may sit at the center of the same supply chain platform, but analysts and executives often evaluate them through entirely different lenses.

Click on the perspective you want to analyze – executives or analysts.

Supply Chain Optimization Self-Assessment

Not every organization uses supply chain optimization in the same way. The following framework can help organizations assess their supply chain optimization maturity.

In Conclusion

Optimization solvers rarely receive the same attention as artificial intelligence, digital twins, or control towers. Yet behind many of the most important supply chain decisions lies an optimization engine evaluating trade-offs that would be impossible to assess manually.

Whether supporting strategic network design, production planning, packaging management, or inventory optimization, solvers provide a structured way to navigate complexity and identify actions that align with business objectives.

At the same time, optimization itself is evolving. Modern platforms increasingly combine exact mathematical methods, heuristics, simulation, and artificial intelligence to solve problems that would have been impractical only a few years ago. The result is not just faster computation, but more realistic decision support.

The organizations that create the greatest value in the future will be those that use each tool where it performs best: human expertise for judgment and oversight, AI for pattern recognition and prediction, and optimization for navigating the complexity in between.

Demand volatility has become a structural feature of modern supply chains, not a temporary disruption to be managed. Weather swings, viral product trends, geopolitical shocks, and sudden shifts in consumer behaviour can reshape demand patterns in days rather than quarters. Traditional forecasting systems, built primarily on historical sales data and updated on weekly or monthly cycles, were not designed for this pace. By the time a conventional model detects a shift, the stockout has already happened or the overstock is already accumulating.

Artificial intelligence is changing that dynamic. By combining machine learning with generative AI, companies can now integrate real-time signals from weather feeds, social media, IoT sensors, and point-of-sale systems and simulate multiple demand scenarios before disruptions materialize. It is a fundamental change in what forecasting means: from predicting a number to understanding demand dynamics as they unfold.

In case you want to jump ahead

1. The Methodologies Behind the Shift

2. Ice Cream as the Case for Real-Time Forecasting

3. Additional Industry Cases

4. Organisational Readiness Check

5. Key Implementation Steps: Choose Your Perspective

6. Conclusion: Sense Faster, Respond Earlier

The Methologies Behind the Shift

To understand what AI is changing in demand planning, it helps to understand what the new methods actually do and how they differ from the statistical models that preceded them. Four approaches currently define the state of the art.

Ice Cream and the Case for Real-Time Forecasting

Few products expose the limits of traditional demand forecasting as clearly as ice cream. And few companies have gone further in rethinking how demand planning works at scale.

Demand for ice cream is simultaneously predictable in aggregate and deeply volatile at the local level. Seasonal patterns are well understood. But a sudden heatwave can trigger a regional surge within 48 hours, while a rainy week suppresses sales just as rapidly. For a product sold almost solely through impulse purchases at the point of a freezer cabinet, with a cold chain that cannot be improvised overnight forecasting errors translate directly into lost revenue, excess inventory, and inefficient logistics.

This is the structural challenge faced by Unilever, whose ice cream supply chain spans 60 countries, 35 factories, and roughly 3 million freezer cabinets worldwide. At that scale, even a 5% improvement in forecast accuracy is worth hundreds of millions of dollars annually. And the cost of a stockout on a hot afternoon when a customer reaches into a cabinet and finds it empty is not just a lost sale. It is a brand interaction that fails at the most critical moment of the purchase journey.

From Historical Extrapolation to Real-Time Demand Sensing

The Freezer Cabinet as a Demand Sensor

The most operationally significant change has happened at the retail edge. More than 100,000 of Unilever’s freezer cabinets are now equipped with AI-powered image recognition systems that monitor stock levels in real time. When a camera detects that a cabinet is running low, the system generates a replenishment recommendation automatically, not based on a projected depletion curve calculated last week, but based on what is actually in the cabinet right now, cross-referenced with the current weather forecast and local sales velocity.

The practical consequence is that roughly 5% of seasonal orders are now generated directly from these AI-driven recommendations rather than from traditional planning cycles. That percentage will grow as coverage expands. For operations teams, this means fewer manual ordering decisions, more efficient delivery routes, and a fundamentally different relationship between demand signal and supply response.

Why Ice Cream Case Matters Beyond Ice Cream

The ice cream case is valuable precisely because it is not a special case. The same pattern applies to beverages, fresh food, seasonal apparel, and any product where the gap between demand signal and supply response carries a real cost.

What Unilever demonstrated is that closing that gap requires three things working together: a real-time data layer (weather, POS, cabinet sensors), a forecasting model capable of processing those signals faster than the planning cycle, and an execution system that can act on the output without waiting for human review. None of those three elements alone is sufficient. The competitive advantage comes from their integration.

The freezer cabinet is not just a storage unit. For Unilever, it has become the primary demand sensor, generating signals that the planning system acts on within hours, not weeks. That is a different model of forecasting.

Additional Industry Cases

The ice cream model is not an isolated experiment. The same underlying shift is playing out across industries, each with its own version of the volatility problem.

Click on the use case you want to learn more about:

Before committing to a technology investment, supply chain leaders should assess their organisation’s readiness across four dimensions. The table below provides a structured framework. Most large organizations entering AI demand planning in 2025–2026 sit at the Developing stage on data and at the Early Stage on governance, the latter being the more consequential gap to close.

The developments described above are not sequential innovations. They are converging simultaneously.

Click on the perspective you want to analyze – executives or analysts.

Conclusion: Sense Faster, Respond Earlier

Traditional forecasting systems were designed to predict numbers. AI-driven systems are designed to understand demand dynamics as they unfold. The combination of machine learning, transformer-based probabilistic models, generative AI scenario simulation, and agentic autonomous execution is transforming demand forecasting from a periodic planning activity into a continuously learning, real-time system.

The Unilever ice cream case makes this concrete. A freezer cabinet that detects its own depletion, cross-references the local weather forecast, and generates a replenishment order without human intervention is not a futuristic scenario. It is already operating at scale across 100,000 retail locations.

For supply chain executives, the strategic implication is clear. Better forecasts improve inventory efficiency, reduce waste, and ensure product availability when demand spikes. But the deeper competitive advantage is structural: organisations that can sense demand faster and respond earlier than their competitors gain a durably superior position.

The organisations that gain structural advantage in the next five years will not be those that deployed AI first. They will be those that built the data foundations, governance frameworks, and organisational capabilities to use it reliably at scale.