Guy F. Courtin - Vice President of Industry and Global Alliances, Tecsys.
At BMW’s Regensburg plant, a vehicle rolls off the assembly line every 57 seconds. Any disruption could cost thousands of dollars in downtime, labor and missed production goals. Yet, thanks to a predictive maintenance system powered by narrow AI, the plant prevents over 500 minutes of downtime annually.
This is just one example of how narrow AI is transforming supply chain operations, turning data into actionable insights and reshaping industries in real time.
From predicting shortages to optimizing delivery routes, it’s driving efficiency and unlocking capabilities that were once out of reach. As these technologies continue to evolve, supply chain leaders must consider how they can harness them to stay competitive.
What is narrow AI?
Narrow AI is designed to solve specific problems by analyzing data within predefined rules, and it improves over time through machine learning (ML). It uncovers patterns and trends that may be difficult or time-consuming for humans to detect, increasing precision with each data-driven adjustment.
Unlike generative AI (GenAI), narrow AI can't “think” abstractly. Similar to chess, narrow AI follows specific rules and predefined functions—you know the rules of engagement, the potential moves and the foundational knowledge. Narrow AI is intended to perform a specific task within set parameters, not to think outside the boundaries of the information provided.
The following are examples of narrow AI:
• AI-powered virtual assistants like Siri and Alexa that respond to voice commands.
• Customer service chatbots that handle inquiries and provide support on websites.
• Predictive text in smartphones that suggests words based on previous usage.
How can narrow AI be used in the supply chain?
Supply chain planners are always on the lookout for new technologies to enhance accuracy. AI’s ability to process vast datasets offers significant potential. Many supply chain AI applications focus on specific challenges, making narrow AI ideal for the following use cases.
Planning
Planning with AI can be tricky. Because AI aims to predict future events, it can struggle when historical data doesn't align with future trends, which creates challenges in forecasting. However, narrow AI can provide improvements in capacity and material planning.
Capacity Planning: The technology is particularly effective in managing tasks with clear, measurable limits, such as capacity management. Supply chain leaders are well aware of their total manufacturing capacity and the constraints that impact operations. By applying narrow AI, companies can develop more accurate models of their capacity. This can help with strategic decision-making.
Material Planning: Narrow AI is proving valuable in material planning. For example, during a conversation with an industry analyst, I learned that Nestlé uses data to predict cocoa crop yields more accurately, which allows them to better plan production. By analyzing historical data, weather patterns and other relevant information, companies such as Nestlé can improve forecasting for raw material yields. The results are better inventory management, optimized production and reduced risks in supply chain planning. These advancements also contribute to more sustainable sourcing and production strategies.
Supply Chain Execution
Narrow AI is increasingly optimizing supply chain execution, from warehouse operations to last-mile delivery.
Warehouse Optimization: Warehouse operations present a set of specific rules, information and outcomes—perfect for narrow AI. A warehouse management system (WMS) acts as a data hub, guiding both workers and automated technologies (e.g., partial truckload systems and cobots) in specific tasks. Narrow AI can analyze data provided by the WMS to offer actionable recommendations, such as optimizing pick paths, slotting and labor assignments to minimize downtime.
Last-Mile Optimization: Narrow AI also has promise in last-mile delivery by optimizing routes and improving efficiency. As demand for faster deliveries grows, AI analyzes past data, fine-tunes routes and reduces fuel use by factoring in real-time conditions like traffic and weather. However, we must continue to value the localized knowledge of delivery drivers, which can't be accounted for within a historical dataset.
It starts with the data.
The performance of any AI engine hinges on the quality of its fuel—and that fuel is data. For narrow AI, the data pool should be intentionally focused on the insights you aim to uncover, the challenges you need to solve or the workflows you want to enhance. This specificity simplifies the process of creating datasets that are clean, relevant and complete. It also makes it easier to maintain their integrity over time.
Start by:
• Establishing a clear understanding of the data you need.
• Developing data policies to compile and shape the required data.
• Bridging gaps between relevant systems to facilitate data capture and sharing.
• Deploying automated data cleansing capabilities to maintain data quality over time.
Turn potential into performance.
With high-quality, relevant datasets, narrow AI becomes more than just a tool. It drives tangible improvements across supply chain operations, from forecasting to optimizing warehouse processes and last-mile delivery.
By leveraging historical and real-time data, narrow AI not only helps improve efficiency but also drives smarter decision-making, better resource utilization and reduced environmental impact. This strategic focus ensures that supply chain leaders are ready to meet evolving demands while fostering a high-performance and sustainable future.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?