Generative AI in the supply chain: from optimisation to anticipation

For years, the focus of artificial intelligence in the supply chain has been on optimisation: reducing costs, improving demand forecasting and automating repetitive tasks. But there is one use of generative AI that is beginning to stand out from the rest and which, until now, has been less prominent: its ability to transform crisis management.

According to the AECOC, 22% of companies already use generative AI in their supply chains, especially in areas such as planning and procurement. But beyond efficiency, its true potential lies in anticipation. Generative AI can simulate complex scenarios and anticipate disruptions, allowing companies to prepare before problems arise. 

From “what if…” to preventative action

Imagine if a company could predict, weeks in advance, how a failure at its main supplier would affect it. Or if it could measure the impact of a strike, a port closure or a global health crisis on its logistics network. Today, these simulations are no longer science fiction. In fact, according to Accenture, 43% of working hours in supply chains could benefit from generative AI.

This has key implications. First, it changes the way companies approach decision-making. It is no longer just about reacting, but about assessing risks and designing customised responses. Second, it redefines the concept of resilience. In an environment where disruptions are no longer exceptional but recurrent, anticipation makes the difference between staying in business or being left out in the cold.

Scenario-based decision-making

Cases such as the COVID-19 pandemic show the impact of being unprepared. Companies that had real-time data and analysis tools were able to react better. Now, generative AI takes this a step further, offering predictive capabilities and much more agile response strategies.

However, the challenge is not solely technological. To leverage generative AI in crisis management, companies must review their data models, ensure information quality and train teams capable of interpreting the data generated. The technology is ready; what is missing is the ability to integrate it intelligently into decision-making.

How can generative AI be applied to manage crises in the supply chain?

1. Collect and standardise key data: generative AI needs data: historical, real-time and from external sources. The first step is to consolidate this information on platforms that allow it to be analysed.
2. Train generative models with real scenarios: use past crisis data (supplier failures, logistics shutdowns, pandemics) to train the AI. This will enable it to generate future scenarios with similar variables.
3. Evaluate responses and adjust strategies: AI not only shows the impact, it also generates potential responses – rerouting, changing suppliers, adjusting inventories. The key is to test them and adapt contingency plans accordingly.
4. Integrate AI into daily decision-making: its use should not be sporadic. Data should feed into dashboards and alerts that enable quick decisions when risks are detected.

The next major transformation of the supply chain will not be solely digital. It will be predictive, dynamic and, above all, resilient. And generative AI can be the catalyst for that change.