Supply chain management changed more through trade policy than through technology between 2024 and 2026. Tariffs rewrote landed costs, artificial intelligence spread widely but shallowly, and the sensors that promised total visibility mostly produced more data than teams could act on. This guide sets out what actually shifted, what the evidence supports, and what a small operations team can do about it. Two points up front: most companies apply AI to narrow tasks rather than redesigning how they plan and buy, and the largest single variable in most 2026 plans is not a technology at all.
Key Takeaways
- Tariffs, not technology, are the dominant planning variable for most companies in 2026.
- AI adoption is broad but shallow: Gartner found 83% of surveyed supply chain leaders applying it incrementally rather than redesigning processes.
- Data quality, not model quality, is the usual reason a supply chain AI project stalls.
- Connected sensors are cheap and plentiful. Turning their output into a decision is the hard part.
- Product traceability is moving from good practice to legal obligation in the EU from 2027.
What Actually Changed Since 2024
The direction of travel has not reversed. Companies still want fewer disconnected spreadsheets, more shared data with suppliers, and a single view of where goods are. What changed is the pressure behind that ambition.
Three forces did most of the work. Tariffs pushed sourcing decisions back onto the executive agenda. Generative AI arrived in planning and procurement tools, mostly as an assistant rather than a replacement for planners. And regulators began asking companies to prove where a product and its materials came from, rather than simply assert it. The result is not a simpler supply chain. It is a more political and better documented one. Our overview of digital transformation trends covers the same period from the wider corporate side.
Trade Policy Is the Biggest Variable in 2026
For most companies, the question that moves the numbers is not which planning tool to buy. It is what a shipment will cost once it clears customs.
What Tariffs Did to Costs and Prices
KPMG surveyed 300 U.S. C-suite leaders at companies with at least $1 billion in annual revenue between February and March 2026. Sourcing costs had risen by more than 25%. A third of respondents (34%) said they now pass more than half of tariff costs on to consumers, up from 13% in May 2025. More than half (55%) planned price increases of up to 15% within six months.
The effects reached beyond pricing: 82% reported declining foreign sales and 61% saw domestic sales fall. For a mid-sized importer the consequence is concrete. Landed cost has to be recalculated per product and per origin country, often several times a year, which is a data problem before it is a strategy problem.
Reshoring Is Slower Than the Headlines Suggest
Moving production closer to the customer sounds like the obvious answer. It is also slow and expensive. In the KPMG survey, 26% of companies were in formal planning or active execution of reshoring, up from 10% six months earlier. But 60% estimated a complete move would take one to three years.
That gap matters. A tariff can change in a quarter; a factory cannot. So most companies hedge instead: qualifying a second supplier in a different country, holding more safety stock on high-risk items, and rewriting contracts so duty costs are shared rather than absorbed. Our guide to supply chain resilience works through those hedges, and cross-border e-commerce covers selling across borders under the new rules.
AI in the Supply Chain Is Broad but Shallow
Almost every planning, procurement and logistics vendor now ships AI features. Far fewer companies have changed how they work because of them.
What the Adoption Data Shows
Gartner surveyed 140 senior supply chain leaders in November 2025 and published the results in May 2026. Only 17% were pursuing an immediate transformational redesign of their processes. The other 83% were applying AI incrementally, to specific use cases, or scaling it gradually.
Gartner’s analysts pointed to a consistent set of blockers: a fragmented vendor landscape with no single platform that does everything, data quality and master data gaps, inconsistent data from supply partners, and a need to upskill staff before the tools pay off. That is not a story about disappointing technology. It is a story about foundations.
Where AI Earns Its Place Today
The use cases that work are narrow and repetitive. Demand forecasting is the clearest: a model that reads point-of-sale data, promotions and weather usually beats a spreadsheet built on last year’s average, particularly for seasonal products.
Document handling is the other reliable win. Purchase orders, invoices, customs paperwork and supplier certificates arrive as PDFs and emails. Software that reads them, extracts the fields and flags mismatches removes hours of manual keying each week in a small purchasing team. Exception handling sits between the two: instead of reviewing every open order, a planner sees the fifty at risk and why. Our overview of AI in business operations covers the same pattern in other functions, and hyperautomation describes what happens when several narrow automations are chained together.
Why Data Quality Decides the Outcome
A forecasting model cannot fix a product master where the same item exists under three codes. It cannot reconcile a supplier who reports lead times in working days with one who reports calendar days. These sound like housekeeping problems. In practice they are why most pilots never leave the pilot stage.
The unglamorous work comes first: agreeing definitions, cleaning the item master, deciding who owns each field. A data governance strategy makes that stick after the initial cleanup, and the data-driven insights everyone wants only hold up once the underlying records agree with each other.
Agentic AI and the Entry-Level Question
Agentic AI, meaning software that plans a sequence of steps and executes them rather than answering a single prompt, was the term vendors used most in 2026. Gartner surveyed 509 supply chain leaders between July and October 2025. A majority (55%) expected it to reduce entry-level hiring needs, and 51% expected shifts toward overall workforce reductions. At the same time, 86% agreed that adopting it will require new ways of developing future talent.
That last figure is the interesting one. Planners learn the trade by doing routine work first: chasing purchase orders, checking stock, running the standard report. If software absorbs that work, companies need a different route to build experienced people. Nobody has solved that yet. The compliance side is covered in our guide to AI regulation.
Robots and Physical AI on the Floor
Physical automation follows a slower curve than software, but the base is large. The International Federation of Robotics recorded 542,000 industrial robots installed worldwide in 2024, the fourth consecutive year above 500,000 units, bringing the operational stock to roughly 4.66 million. Asia accounted for 74% of installations, with China alone taking 295,000 units.
Gartner’s list of top supply chain technology trends for 2026, published in June 2026, shows where this is heading. Polyfunctional robots handle several tasks rather than one fixed job, which lowers the volume threshold at which automation makes sense. Physical AI links models to sensors and robots so decisions get executed on the floor rather than recommended in a report.
For a warehouse handling a few thousand orders a week, the question is narrower: does a mobile robot that moves shelves to a picker pay back against the cost of hiring? That depends heavily on local labor availability. Our guides to robotics automation, predictive maintenance and private 5G networks in manufacturing cover the operational and infrastructure side.
IoT and the Limits of Visibility
The Internet of Things, meaning physical objects with sensors that report their state over a network, is no longer emerging. IoT Analytics counted 21.1 billion connected IoT devices worldwide in 2025, up 14% year on year.
What Connected Sensors Actually Deliver
In a supply chain the useful cases are concrete. A temperature logger in a refrigerated container proves a cold chain held, which settles insurance claims faster. A GPS unit on a trailer gives a customer an honest arrival window instead of a guess. RFID tags at a warehouse door reconcile a received pallet against the advance shipping notice without anyone counting boxes.
Each settles a specific dispute or removes a manual task, which is why they get funded. Our wider guide to IoT in business operations covers the same pattern outside logistics.
Visibility Is Not the Same as a Decision
Here is the trap. Knowing where every pallet is does not tell you what to do when one is late. Someone still has to decide whether to expedite, substitute, or let the customer wait. Gartner surveyed 151 supply chain leaders at companies with at least $250 million in revenue between November and December 2025. It found 72% had to revisit final approvals for network decisions at least once, which added delay, and more than half revisited them three or more times. The bottleneck was not missing data. It was an unclear decision process.
So before buying another tracking layer, write down who decides what, at which threshold and with which authority. That is cheaper than a platform and usually fixes more.
Traceability Becomes a Legal Requirement
Proving provenance used to be a marketing exercise. In the EU it is becoming a compliance one. The Ecodesign for Sustainable Products Regulation (EU) 2024/1781 introduces the Digital Product Passport, a digital record holding sustainability, composition and compliance information for a product.
The timeline is public. The DPP Registry became operational on 20 July 2026. From 18 February 2027 the passport is mandatory for certain battery types, including electric vehicle, light transport and industrial batteries. Textiles, aluminium and tyres follow in the second half of 2027, with furniture, mattresses and other groups from 2028.
If you sell physical goods into the EU, you will need supplier data you probably do not collect today: material composition, origin and recycled content, item by item. Gartner lists product provenance among its 2026 trends for that reason. The technology follows the requirement, whether that is blockchain in logistics or a better supplier portal.
Cloud Platforms and the Integration Problem
Supply chain software moved to the cloud years ago. The current problem is not hosting. It is that a typical mid-sized company runs an ERP, a warehouse system, a transport system, a procurement tool and several supplier portals that do not agree with each other. That fragmentation is exactly what Gartner named as the leading barrier to scaling AI in supply chains.
Two rules help. Pick the system of record for each data object and stop letting the others overwrite it. And when evaluating a new tool, weigh its integration options as heavily as its features. See our overview of cloud computing trends and our guide to digital procurement for the buying process.
What to Do in the Next Twelve Months
If you run or influence a supply chain in a small or mid-sized company, the useful moves are unglamorous:
- Recalculate landed cost by product and origin, and redo it whenever duty rates change.
- Qualify a second source for the ten items that would hurt most if they stopped arriving.
- Clean the item master and agree lead time definitions with your top suppliers before buying any AI tool.
- Write down who decides on expedites, substitutions and allocation, and at what threshold.
- Ask suppliers now for the material and origin data the EU passport rules will require.
- Pick one narrow AI use case with a measurable baseline, such as forecast accuracy on your top 50 products, and judge it on that number.
Analysts will keep publishing trend lists. The companies that benefit are the ones that fixed their data and their decision rights first. For the wider picture see our piece on the future of business analytics and on e-commerce trends, which shows what customer expectations now put on fulfilment.
Found this useful?
Make SmartKeys a preferred source on Google, and our articles will surface more often in your Top Stories, AI Overviews, and AI Mode.
Add as Preferred Source
