Increased collaboration between AI and human planners

Supply chain planners have the opportunity to change the forecasts generated by self-learning systems using artificial intelligence. Does this add value? And which forecast can be relied upon more—software or human? In a modern study, KLU researchers Naghmeh Khosrowabadi, Professor Kai Hoberg and Professor Christina Imdahl (Eindhoven University of Technology) are getting to the bottom of these questions. According to their study, human intervention—on average—did not increase the accuracy of the predictions.

For the study, the team analyzed data of 30 million forecast per product unit, per store per day from the leading AI supplier and major grocery retailer in Europe.

Planners’ optimism leads to more inaccurate forecasts

The results show that planners—on average—do not contribute to forecast accuracy. Instead, planners even tend to overcompensate for effects like weather or price drops that are already considered by the AI ​​system, says Khosrowabadi. In the study, only 50 percent of human interventions resulted in improved outcomes.

A closer look at the data also shows that around 5% of AI-generated forecasts have been adjusted by supply chain planners. “We wanted to know why decision-makers decided to adjust AI-generated forecasts,” explains Naghmeh Khosrowabadi. “Our results suggest that product characteristics such as price, novelty, or price reduction are key drivers for the frequency of planners’ adjustments to AI forecasting.”

For example, if the AI ​​system is making a forecast for a very expensive product, planners tend to pay more attention and are more likely to intervene, such as by adjusting the forecast. “Besides, our results show that a big boost from AI forecasting, such as when human forecasting of items will go on sale on a certain day in a particular store is large. more than twice as often as AI forecasts, which happen more often but are also often inaccurate. Over-optimism on the part of planners seems to be the problem here,” said Professor Kai Hoberg. On the other hand, the drop from the AI ​​forecast is less likely but more accurate.

Improve collaboration of human and AI planners

“Human planners will continue to play an important role in AI-powered forecasting processes,” said Professor Hoberg, to increase the odds of better forecasting by more than 70%. That’s why we need to strengthen the collaboration of planners and AI.”

For this, the team recommends more communication between retailers and AI vendors: the better planners understand how the system makes forecasts, the easier it is for them to decide when to intervene. “Using the results of our research, companies can save money and time,” Khosrowabadi promises, “It’s important to help planners decide when they need to intervene—and when the system stabilizes itself and they can focus on other tasks.”

The study involved Naghmeh Khosrowabadi in her doctoral thesis, Professor Dr. Kai Hoberg and Professor Dr. Christina Imdahl from KLU. They analyzed data from 30 million store-level forecast SKUs for the day from a leading AI vendor and a major European retailer. Data on additional variables such as products, weather or holidays are also considered.

The work is published in the magazine European Journal of Operations Research.

Provided by Kühne University of Logistics