Sign up


Prescriptive AI now far outpacing predictive supply chain systems

Turin, March 28, 2021

by Fabrizio Fantini

AI has promised transformation through more accurate demand forecasting. Supply chain efficiency gains were expected to come from better predictions of customer behaviour.

Over the past year, however, it’s become clear that predictive analytics is not as effective as hoped. Gains are limited by the inherent uncertainties surrounding customer choice and market conditions. When significant disturbances, such as a pandemic, occur, predictive models can’t efficiently adjust. Instead, prescriptive AI is emerging as a fully mature solution.

Prescriptive AI systems are learning machines explicitly built to achieve a particular objective. Unlike predictive models, which can only tell you what is likely to happen, assuming no major disruptions, prescriptive models use data to show you how to achieve that goal best. When unexpected events inevitably occur or customer preferences shift, the prescriptive system can more adeptly adjust.

During the early stages of the Covid-19 pandemic, for example, many companies using predictive AI saw their models fall apart or give useless recommendations. Those that continued to deliver actionable advice and allowed companies to pivot successfully to reduce the impact of the crisis? Prescriptive models.

In practice, this agility makes prescriptive systems more responsive and deliver greater ROI. Why? Because, by itself, even a perfect forecast could not produce optimal profit-driven decisions. Pricing curves, brand exclusivity, and other business considerations greatly influence the ideal number of items to ship, regardless of demand. A predictive system designed to forecast sales cannot accommodate these nuances. Only a prescriptive approach builds in all the necessary considerations intentionally.

The difference in results is significant. While predictive AI can only increase demand forecast accuracy so much, prescriptive AI regularly increases supply chain efficiency by 20-30%. In a case study published by Microsoft, Boggi Milano increased sales by 4% and inventory efficiency by 18% in just a few weeks after switching to a prescriptive model developed by AI experts at Evo— and as with all machine learning systems, these results are likely to improve as the model matures.

Predictive analytics are unlikely to disappear; they are still critical to many operations and highly efficient in certain functions. However, prescriptive models will continue to grow in importance. Any business looking to optimize its performance with AI should consider a prescriptive system. According to researchers, it is the future of AI.

To learn more: To Forecast or Not to Forecast, That Is the Supply Chain Question

Hey! Was this page helpful?

We’re always looking to make our docs better, please let us know if you have any
suggestions or advice about what’s working and what’s not!

Send Feedback