The Demand Forecasting Challenge: Taking a Bite Out of the Supply Chain with AI

Staying in tune with consumer demands can be challenging, especially when these demands change almost daily. Tracking consumption patterns to predict demand is a monumental task. Previously, solving this problem involved spreadsheets or legacy statistical methods.

But with automation and machine learning, artificial intelligence can help organizations slice through mounds of data to get more accurate and timely insights on consumer demand. Machine learning can build forecasts for hundreds of thousands of items and consider all of the minute details for each one of them, from seasonality to sales history. For example, automated machine learning can easily create separate models on clusters of products, using unsupervised ML techniques, to make predictions even more granular such as by store, week, and SKU.

One of DataRobot’s customers, Harris Farms, was incredibly hard-pressed to deliver a comprehensive, trusted, and scalable AI solution that could help forecast demand around perishable items in their stock. 

Their specific situation was particularly challenging since they were dealing with extremely anomalous supply and demand patterns. First, wildfires ravaged Australia, making their inventory management harder. Then with COVID, they started experiencing demand spikes. Fresh produce is roughly 50% of their business, so taking control of this category’s supply chain meant bringing the much-needed stability to this retailer in especially turbulent times.

Their Head of IT, Phil Cribb, recognized the value of automation and AI in their quest to deliver better supply chain predictability.

After just a few months, Harris Farms utilized DataRobot to deliver around a hundred models, with a primary focus on the fresh produce SKUs in their inventory. They are currently deploying 400 forecasts around a wide range of products, and DataRobot serves as a platform that delivers trusted information for decision-making around the chain’s stocking operations.

And the impact for retailers, like Harris Farms, doesn’t stop there. Some of the most valuable AI use cases in retail include: 

• Path-to-purchase predictions
• Returns forecasting
• Promotional optimization
• Site optimization and best future store location predictions

For more details around use cases and potential applications of AI in retail, check out our retail industry overview. This ebook takes a deeper look at the challenges facing the marketplace and how AI can help retailers put strategic solutions into place that will help them grow and refine their businesses in a fast moving world.

About the author

Ari Kaplan

AI Evangelist, DataRobot

Kaplan is a leading figure in data science, sports analytics, and business leadership. High profile roles include creating the Chicago Cubs analytics department, President of the investigation into the fate of Holocaust hero Raoul Wallenberg, and President Emeritus of the worldwide Oracle User Group.

Meet Ari Kaplan