1.4 Million Units in Six Months for an AI-Selected Product
CJ CheilJedang's 'Bibigo Salmon Steak,' planned based on the company's proprietary AI platform 'Food AI 360,' surpassed 1.4 million units sold within six months of launch. In June alone, 265,000 units were sold, equating to roughly one unit sold every 10 seconds. The product reflects a shift in protein consumption trends toward white meat like chicken and seafood-based protein, as well as diversifying consumer preferences moving from traditional side-dish mackerel toward salmon, which better fits diet-management demand. 'Matcha Hetbahn' is another example born from Food AI 360, which confirmed the sustainability of the matcha trend that has continued since last year and translated the consumer demand for healthy ingredients combined with convenience into an actual product.
Not Just Another Tool, But Integrated Steps
Food AI 360 is an AI-based marketing solution that upgrades the existing workflow—from trend discovery to concept development, evaluation, and market response monitoring—into a single integrated platform. CJ CheilJedang first introduced Food AI 360 to its domestic business division in June of last year, and selected the U.S. as its second market for global K-food expansion, completing the rollout there this past June. A CJ CheilJedang representative emphasized, "Food AI 360 is a platform that connects the entire process—from trend detection to consumer evaluation and commercialization—through AI, simultaneously accelerating development speed and increasing the likelihood of market success."
What Was Already Built Outside of Marketing
AI adoption isn't limited to product planning. CJ CheilJedang became the first in Korea's food industry to introduce an MI Room in 2019, and has since expanded AI applications within it to monitor in real time not only internationally traded raw materials such as raw sugar, wheat, and soybeans, but also domestic agricultural products, exchange rates, and oil prices. In 2023, when futures market prices rose amid fears over El Niño, the company delayed its raw material procurement timing to avoid buying at peak prices. On the production floor, AI deep learning based on X-ray imaging was first introduced in 2024 at the Jincheon (Chungbuk), Incheon, and Yeongdeungpo (Seoul) plants for automated defect detection in quality inspection, and has since been expanded further. At its sugar production plants, AI now controls the amount of steam used in the crystallization process that converts concentrated sugar solution into crystal form, reducing average annual steam usage by about 6% in 2024 compared to 2023.
What to Count First Within Your Own Organization
What matters in this case is the sequence, not just the outcome. A centralized room for monitoring raw material market conditions came first, in 2019. Deep learning was then applied to tasks like inspection and process control, where right and wrong can be measured numerically. Only after that were the previously scattered steps—from trend discovery to commercialization—unified into a single platform. For organizations considering adoption, the right starting point is to count how often a particular decision recurs each week and what data informs it at that moment. If your organization already has a metric being tracked, such as steam usage, it's worth checking whether that metric can itself serve as proof of the impact of adoption, before and after.
Source: AI Handles Everything from Food Trend Analysis to Commercialization... Giving Wings to K-Food