South KoreaFood and restaurants2025 – ongoingLotte Chilsung Beverage
AILISS generative AI product-label review
Lotte Chilsung reported 95% character-recognition accuracy in AI product-label review
Problem
Employees manually checked product-label information against product records and food-safety information. Complex label fonts made extracting text difficult.
Approach
Lotte Chilsung built AILISS to extract and review label information. OCR read label images, while generative AI classified the extracted content and agents checked errors against supporting records. A retrieval-based chatbot answered questions about relevant regulations.
Results
Takeaway
The reported accuracy covered character recognition. It did not measure the accuracy of the complete regulatory review.
Categories
- Size
- Large enterprise
- Industry
- Food and restaurants
- Business function
- Production and qualityFinance, risk and legal
- Innovation type
- Process innovation
- AI technology
- Generative AIAI agentsComputer vision
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Lotte Chilsung had built AILISS to automate product-label checks using OCR, AI agents and retrieval-augmented generation. The report gave 95% character-recognition accuracy for complex fonts and described checks against product and regulatory information.
Full details are in the original sources.