South KoreaRetail and e-commerce2026 – ongoingBGF Retail (CU)
AI cooking-quantity forecasts for in-store fried chicken
CU's AI cooking forecast helped one store sell four times its usual chicken on a World Cup match day with under 1% waste
Problem
Convenience stores cook fried chicken in store. On event days such as football matches demand jumps, and store owners guessing quantities risk both selling out and throwing food away.
Approach
CU's machine-learning forecasting system calculates suitable cooking quantities for about 7,000 stores. It learns from each store's sales over the past year, combines the average of the last four weeks with a seasonal index and changes in the local trade area and customer mix, and gives each store a report recommending quantities by time slot and item.
Results
Takeaway
The published result covered one store on one day; no chain-wide waste or sales figures have been released. The system forecast event-day demand that store owners had previously judged by feel.
Categories
- Size
- Large enterprise
- Industry
- Retail and e-commerce
- Business function
- Logistics and supply chainProduction and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
The Korea Economic Daily reports that on 12 June 2026, the day of a World Cup qualifier, a CU store in Seoul's Gangnam district sold about four times its usual chicken with a waste rate under 1%, guided by a store-level AI report. It describes the forecasting method and says the system covers about 7,000 stores. A CU official said store profitability improved greatly, without giving a figure.
Full details are in the original sources.