In this article, learn about:
How traditional forecasting struggles in times of tariff volatility
How integrating AI forecasting can help
The data foundations essential for successful AI forecasting
How SPS Commerce can set suppliers up for success with AI forecasting
Summary: AI demand forecasting is giving enterprise CPG brands a critical edge against tariff-driven supply chain disruption. By modeling cost-impact scenarios, integrating real-time trade signals, and optimizing inventory before policy shifts take hold, leading brands are moving from reactive cost containment to proactive margin protection.
Introduction
2025 may go down as the year that finally broke static supply chain planning in the consumer packaged goods (CPG) industry. Blanket tariffs, country-specific levies, and retaliatory trade actions cascaded through global supply chains with a speed that left demand planners scrambling. Consumer-facing companies projected a combined financial impact of more than $21 billion from tariff disruptions in 2025 alone. For CPG brands selling through major retailers, that pressure did not stay in the sourcing department. It traveled straight through the P&L.
Brands navigating this environment with the least disruption share a common capability: AI demand forecasting. Not AI as a buzzword attached to a legacy planning tool, but genuine machine learning applied to the full picture of demand signals, cost structures, and trade risk. This article breaks down how leading CPG companies are deploying these tools and what separates the brands getting results from those still reacting after the fact.
Why Are Tariffs So Disruptive to CPG Supply Chains?
Tariffs are not simply a cost line. A single policy shift can upend sourcing strategies, destabilize supplier relationships, and force costly recalculations across global landed cost models, as Jabil's supply chain analysts describe. The landed cost, the full expense of getting a product to its delivery point after accounting for duties, freight, and compliance overhead, can shift overnight without any change to the underlying demand environment.
That mismatch is what makes tariff volatility uniquely dangerous. CPG brands carry promotional commitments, retailer OTIF (on-time, in-full) obligations, and seasonal inventory positions built on cost assumptions that may no longer hold. When those assumptions collapse, the options are ugly:
Absorb margin erosion
Pass costs to retail partners and risk pushing buyers toward private label
Scramble to source alternates under tight timelines
A KPMG survey of 300 C-suite executives in 2025 found that 77% of consumer goods companies had renegotiated supplier contracts in response to changing tariffs, the highest rate of any sector surveyed. That figure reflects how broadly tariff volatility is reshaping sourcing decisions. But renegotiation is a downstream response. The brands gaining a structural edge are the ones identifying exposure before it becomes a crisis.
CPG supply chains are also uniquely prone to the bullwhip effect, the pattern in which small demand fluctuations amplify into costly overreactions at each tier of the supply chain. Tariff-driven demand uncertainty accelerates this dynamic. When planners cannot confidently project how a price increase will affect sell-through at the store level, they build in safety buffers that cascade upstream. The result is excess inventory in some nodes, shortages in others, and margin erosion throughout.
What Does Traditional Demand Forecasting Get Wrong?
Most CPG demand planning still relies on historical sales data, fixed seasonality models, and planning cycles that refresh monthly or quarterly. These tools were built for a more stable world. They tend to smooth over volatility rather than embrace it, making them slow to react to events such as supply disruptions, sudden price increases, or shifts in consumer behavior.
The failure mode in a tariff environment is specific. When a tariff increases landed cost and forces a price adjustment at retail, consumer response is not linear. Price-sensitive shoppers may switch to a lower-cost alternative, a competitor's product, or a store brand. A traditional forecasting model would have no mechanism to anticipate that behavioral shift because it was not present in historical data. It would continue projecting demand based on prior sales patterns, leading to overproduction, excess inventory, and eventual markdown pressure.
A June 2025 Gartner report found that only 23% of supply chain leaders had a formal supply chain AI strategy in place within their organizations. Most were pursuing short-term, unstructured projects rather than the long-term transformation required to make AI effective. That gap between AI adoption and AI operationalization is where the real competitive advantage lies.
How Does AI Demand Forecasting Help CPG Brands Navigate Tariff Volatility?
AI demand forecasting does not simply produce more accurate numbers. It changes what questions a planning team can ask, and how quickly they can get credible answers. Below are the specific capabilities enterprise CPG brands are using to manage tariff risk.
Scenario Modeling Before Policy Takes Effect
Leading CPG companies use AI to run what-if scenarios against proposed or anticipated tariff changes before they go into effect. A model can simulate the downstream demand impact of a 15% price increase at retail across different customer segments, channels, and regions, allowing supply chain and commercial teams to align on a response strategy rather than react to one.
This kind of scenario modeling has historically required weeks of manual analysis across finance, sales, and operations. AI-powered planning tools compress that timeline to hours, giving CPG brands a decision window that simply did not exist under traditional planning approaches. AlixPartners recommended this posture explicitly for CPG leaders in early 2025: harden supply chain strategy, conduct thorough scenario modeling upfront, and develop risk and resilience plans before major disruptions occur.
Real-Time Signal Integration
Traditional forecasting models ingest internal historical sales data. AI demand forecasting models integrate external signals, including point-of-sale (POS) data from retail partners, macroeconomic indicators, competitive promotional activity, weather patterns, and trade policy feeds. When a tariff is announced, the system updates its assumptions across all of those inputs simultaneously rather than waiting for the next planning cycle.
When a tariff-driven price increase causes consumers to switch brands, an adaptive AI platform can flag the sales pattern within days and allow teams to adjust promotions or pricing before the revenue impact compounds. Static models won’t register the behavioral shift until it shows up in lagging historical data, weeks later.
For suppliers sending advance ship notices (ASNs) to retail partners, this real-time visibility matters beyond the warehouse floor. Retailers are watching supplier OTIF rates closely, and supply chain disruptions that follow tariff-driven sourcing changes create compliance risk. AI systems that can anticipate disruptions allow teams to communicate proactively rather than explain shortfalls after the fact.
Inventory Optimization and Safety Stock Recalibration
Companies using AI-driven demand planning have reported a 20 to 30% reduction in inventory costs and up to a 65% improvement in forecast accuracy, according to Gartner and BCG studies. For CPG brands managing hundreds or thousands of SKUs across multiple retail partners, those gains translate directly to working capital released from excess stock and fewer emergency procurement decisions made under pressure.
Safety stock levels are a particular challenge in a tariff environment. The instinct is to buffer aggressively, holding more inventory ahead of potential cost increases or supply disruption. AI forecasting can calibrate those buffers more precisely, holding enough inventory to protect service levels without tying up capital in product that may face markdowns if demand forecasts shift. That precision matters especially for brands with perishable or seasonal products where excess inventory cannot be carried indefinitely.
What Does AI for CPG Actually Require to Work?
The limitations of AI demand forecasting are data limitations. An AI model is only as accurate as the data feeding it. Operations leaders cite data issues as a top reason that technology investments have not delivered expected results. The challenge is not primarily internal. The most valuable supply chain information is the data that flows between trading partners, and many suppliers still exchange critical information manually.
Purchase orders, inventory updates, and fulfillment signals that move through email or spreadsheets cannot be ingested by AI systems in real time. They arrive late, require manual reconciliation, and introduce the kind of inconsistency that degrades model performance. For AI demand forecasting to work at scale, the data exchange between a CPG supplier and its retail partners needs to be automated, standardized, and timely.
McKinsey's 2024 survey found that 71% of CPG leaders had adopted AI in at least one function, up from 42% the prior year. But adoption rates do not reflect operational effectiveness. The gap between having AI tools and having the data infrastructure to make those tools perform is where most implementations stall.
AI for supply chain optimization requires starting with the problem, not the technology. Supply chain leaders who have seen meaningful results from AI projects anchored them to specific, measurable outcomes: improving forecast accuracy in a specific category, reducing stockouts in a key retailer, or flagging cost-exposure windows earlier in the planning cycle. Broad AI initiatives without clear success criteria rarely survive contact with real operational complexity.
How Does the SPS Commerce Network Enable AI Demand Forecasting?
The data foundation that AI demand forecasting requires is not built by any single company in isolation. It is a product of connected, standardized data across a supplier's entire retail network. That means purchase orders arriving electronically, inventory positions shared automatically, and sales data flowing from retailer to supplier without manual intervention.
SPS Commerce Fulfillment automates the order management workflows that create this foundation. By standardizing how data moves between CPG suppliers and their retail partners, SPS Commerce ensures that the information flowing into AI forecasting systems is clean, timely, and consistent. Across a network connecting more than 120,000 trading partners, that standardized data exchange is what separates supply chains where AI performs from supply chains where it struggles.
When tariff conditions shift and demand signals change, AI systems need partner data, including lead times, capacity updates, and inventory positions, to recalibrate quickly. If that data is delayed or inconsistent, the AI recommendations go off course. A system might call for higher order volumes at exactly the moment a supplier is at capacity or assume stock availability that does not exist. Automated, reliable data exchange eliminates those issues.
SPS Commerce Analytics gives suppliers direct access to retailer POS data to fuel demand planning. Rather than building forecasts from internal shipment data alone, suppliers can see actual sell-through rates at the store and regional level, the same signal that AI forecasting models need to detect demand shifts early. In a tariff environment where consumer response to price changes varies by market and channel, that granularity matters.
What Should CPG Brands Do Next?
Tariff volatility is not resolving quickly. Under the Trump administration, tariff and trade policies have continued to shift with little warning, creating ongoing uncertainty across the supply chain. For CPG brands still relying on static planning cycles and manual data exchange with retail partners, each policy shift is a fire drill. For brands with automated data infrastructure and AI demand forecasting in place, it is a planning input.
The difference between those two positions is not primarily about which AI vendor a brand has selected. It is about whether the underlying data infrastructure can support AI at all. Suppliers who have not yet automated their trading partner data exchange are not ready to operationalize AI forecasting, regardless of the tools they deploy on top.
The starting point is operational data connectivity:
Automated order management
Standardized EDI (electronic data interchange) document exchange
Real-time POS visibility from retail partners
Those capabilities create the foundation that AI supply chain solutions need to perform. From there, AI can help CPG brands do what no traditional planning process could: model trade risk before it materializes, detect demand shifts as they happen, and make inventory decisions with the full picture rather than a lagging summary of it.
Building a Data Foundation for AI with SPS Commerce
Ready to build the data foundation your AI strategy requires? SPS Commerce Fulfillment automates order management and trading partner connectivity across your full retail network. See how leading CPG brands are using SPS Commerce to power smarter forecasting and stronger retailer relationships.
