Carrefour turns to AI for demand forecasting
- Carrefour announced the selection of SAS Viya, an artificial intelligence-supported analytics platform, as a way to gain insight from data generated in its stores, warehouses and online orders, according to a press release Monday.
- The French grocer, with more than 12,200 stores in over 30 countries, will use the AI system to help avoid waste and out-of-stocks, optimizing the supply chain and freeing up teams to focus on differentiated forecasting strategies.
- The company will deploy the SAS AI technology in stores after an 18-month test period. Once implemented, it will allow Carrefour to create algorithms specifically for its unique demand forecasting needs, the release said.
Demand forecasting is nothing new. It's been done for decades with more traditional methods like regression, but these techniques sometimes struggled to produce the desired result. A Harvard Business Review article from 1988 put it more bluntly: "Some even throw up their hands and assume that business planning must proceed without good demand forecasts."
Now, technology companies like SAS are saying AI will be able to use the large amounts of data created by supply chains to improve the process of demand forecasting.
In the past, retailers have often relied on overstocking to prevent going out-of-stock on an item. But this can cost the company money by buying products that may not be sold, according to Symphony Retail, which also creates AI products for the retail market.
Traditional demand forecasting systems ingest data, apply it to a set of rules for analysis, and generate a result; but machine learning allows the system to become adaptive, responding to changes in the data it is fed and updating results, Symphony Retail said in a report.
Carrefour is not the only company looking to emerging technology to improve its forecasting. Nike recently pointed out the importance of demand forecasting in an earnings call, with company CFO Andy Campion saying it could increase their margins.
Even the vending machine community is getting in on the trend. A 2018 report from PSFK, a retail intelligence company, and Microsoft highlighted how the Mars Drinks company is using AI to improve its vending machine stock.
Mars is using remote sensor technology from Microsoft and predictive computing to keep up stock, track customer interactions and account for contextual factors that affect demand, such as weather or holidays, according to the report.
Follow Matt Leonard on Twitter