From new tariffs and trade uncertainty to geopolitical tension and extreme weather events, external forces have upended traditional demand forecasting approaches. Among those most impacted are the CPG and retail industries, which have always been more vulnerable due to shifting consumer behaviors, fast-moving products, and complex global supply chains. And while those sectors have been among the first to modernize and onboard new technologies, a considerable gap has emerged between acquiring AI capabilities and effectively operationalizing them. 

A June report from Gartner revealed that just 23% of supply chain leaders had a formal supply chain AI strategy in place within their organizations. The report found that most chief supply chain officers (CSCOs) were instead taking an unstructured approach, focusing on short-term projects instead of long-term, transformational strategies. That’s partly to do with an intensified pressure on C-suite leaders to show quick ROI but, while use cases can be helpful in times of stability, those days are long gone. 

Model Drift: The Silent Killer of Accuracy 

Supply chain leaders are always on the lookout for model drift, when algorithms built on historical data lose their predictive power due to shifting external conditions. Those conditions are becoming increasingly erratic, and the impacts increasingly serious. Excess inventory can jeopardize working capital, stockouts can erode consumer trust, and inaccurate forecasts can result in poor pricing strategies that can tank margins. According to McKinsey, model drift is among the top reasons AI implementations fail to scale effectively across enterprises. 

Unexpected tariffs have been big drift accelerators this past year, and companies such as Walmart and Costco have faced tariff whiplash. When costs swing overnight, supply chains just don’t have enough time to recalibrate, resulting in empty shelves or surplus stock. The situation is even more dire for those businesses reliant on imported goods, where tariffs can increase retail prices by 3% to 5%. Static forecasting models are unable to keep pace with such a dynamic, unpredictable world, and drift has become a constant risk requiring constant vigilance. 

The good news is that adaptive AI/ML capabilities have been built for this moment. But for retailers to thrive, they must reimagine demand forecasting across three critical dimensions — signals, models and collaboration. 

Three Key Levers of Self-Correcting Demand Forecasting 

Better signals: Seeing the unseen. Most traditional forecasting tools rely heavily on structured internal data (i.e. restricted to what’s available in their enterprise) — sales history, POS trends, inventory levels, among others. But there is also a whole world of external variables that can influence how demand shifts, and how consumers buy products. It’s these unstructured signals that are often the first indicators of trouble, and it’s an area that supply chain leaders are actively addressing with applied AI. 

For example, let’s say a retail distributor in the UAE noticed a puzzling sales slump one month. Usual event calendars showed nothing, until analysts conducted deeper research and found regional flooding likely caused product shortages. Traditional models would have missed it, and modeling for one-off events is notoriously difficult. Future forecasting could use AI agents to detect anomalies and investigate disruptions automatically, enabling faster, more accurate, region-specific predictions. 

Better models: Continuous, contextual AI. Even advanced machine learning models degrade over time if they’re not modified or retrained. Adaptive AI models, like agentic AI, continuously learn from real-time data, making them resilient to drift and more aligned with market dynamics. 

To use a hypothetical scenario, consider a beverage brand that faced a sudden tariff-induced hit that pushed its flagship drink price from $8 to $12, causing price-conscious customers to switch to a competitor’s $10 alternative. While traditional models wouldn’t account for this behavioral pivot, an adaptive AI platform would have flagged the sales drop within days, allowing the brand to adjust promotions and save revenue. It’s that sort of robust modeling that is necessary nowadays. 

Better collaboration: GenAI-driven consensus forecasting and linking RGM. Other major failures in current demand forecasting are due to organizational fragmentation and outdated consensus forecasting. Sales, marketing, operations, and finance teams often operate in silos, creating conflicting forecasts that dilute execution. And even when all these various signals are successfully combined into one model, you still need a bunch of planners from each department to sit across a table and agree on a path forward. It’s not only an operational challenge; it can also be a clash of personalities. 

However, by using GenAI-driven simulations, a company can reduce its planning cycle from 30 days to just seven, aligning its price, promotion, and supply scenarios on a shared forecasting platform. Not only does this save time and increase forecast accuracy; it allows for faster responsiveness, and can dramatically improve internal collaboration. It’s also worth noting that demand forecasting influences (and is influenced by) Revenue Growth Management (RGM), especially in areas like pricing, discounting, and promotional strategies. When RGM and demand forecasting co-evolve, margin outcomes improve dramatically, and companies that link the two unlock a powerful feedback loop. 

Wide-Ranging Applicability 

Demand forecasting optimization doesn’t just apply to retail and CPG companies. AI-driven innovations have broad cross-industry relevance, including: 

Manufacturing. Predictive maintenance models improve when upstream retail demand signals are factored in, allowing production schedules to better match market reality. 

Automotive. Tracking real-time consumer interest shifts, economic trends and weather events can help improve dealership inventory. 

Industrials. Companies that produce tangible goods for manufacturing, construction, infrastructure, and defense systems are already benefiting from AI-powered demand sensing, especially when combining IoT, weather and macroeconomic data into their planning platforms. 

The CPG and retail industries can no longer afford to rely on fragile, backward-looking forecasting systems. With adaptive AI, supply chain leaders can get ahead of problems by detecting anomalies before they develop further, updating models as market conditions evolve, and aligning departments on shared, data-driven decisions. In a world of constant disruption, flexibility is quickly becoming the new standard, and only the most resilient will survive. 

Sunder Balakrishnan is director of supply chain analytics at LatentView Analytics.