[NE Coffee Chat] AI Analyzing My Taste Recommends Food... Pamkit 'FoodQ'

Interview with Yang Sunheung, CEO of Pamkit
Product Recommendations Perfectly Matching Individual Tastes and Preferences
Providing an Experience Comparable to Large Shopping Malls Even in Food Specialty Malls

Online video services (OTT) recommend content tailored to viewers based on their viewing history and other factors. Even while wondering how it knew what I liked, I naturally end up clicking. What if the same approach were applied when purchasing food online, exposing products that perfectly match an individual's taste and preferences? Could this increase both consumer satisfaction and seller sales? Farmkit is a startup that began in 2020 based on this question. The technology analyzing each person's unique food preferences is artificial intelligence (AI).

On the 29th, Yang Seonheung, CEO of Farmkit, said, "The purchase conversion rate in the food e-commerce sector is usually around 4-5%, but after adopting Farmkit's solution, it has been shown to rise to about 10%." Showing food products customized to individual tastes more than doubled the rate of actual purchases. It may be obvious that recommending products likely to be liked increases the chance of purchase, but the technology required to achieve this is not simple. Farmkit's solution, FoodQ, analyzes consumers' food preferences through purchase history, clicks, shopping behavior, and reviews. It then adds detailed analysis of product data and provides hyper-personalized recommendations based on time, situation, and location. CEO Yang explained, "Just like product displays in offline stores, which products are shown on the first screen online determine sales. We increase satisfaction by exposing products that are appropriate not only to consumers' tastes but also from a nutritional perspective."

Yang Seonheung, CEO of Farmkit, is explaining Food AI technology.

Farmkit's AI does not simply recommend products similar to those the consumer has shown interest in. For example, someone who likes chicken might eat fried chicken but not baeksuk (boiled chicken). Yang said, "Taste preferences are determined by various factors such as texture, flavor, and raw ingredients, which we have organized into 150 attribute categories." They improved not only accuracy but also speed. Members are analyzed based on purchase history, and even non-members' preferences are quickly identified by analyzing their 'shopping journey' in real time.

Another advantage is the economical implementation cost compared to competing solutions. Global companies with similar functions to FoodQ offer generic solutions that are not specialized for food and are about ten times more expensive. Yang said, "Tests showed that FoodQ's purchase conversion indicators were 60% higher. The process is automated, so it can be easily adopted without expert knowledge of AI, and the subscription model keeps initial costs low."

Yang's personal experience also influenced the development of FoodQ. While working on AI new business strategies at Hyundai Motor Company, he observed that as cars offered more colors and options, consumer preferences became clearer. However, as the number of choices increased, the purchasing process became more difficult. Seeing that AI could help, he decided to apply it to the food sector, which he had long been interested in. Yang said, "Online food purchases are increasing, but customers cannot sample products and it is difficult to read all reviews. In a situation where differentiating customer experience is important, I believed there would be a strong need for personalized food recommendations."

Farmkit's target customers are specialty malls that find it difficult to implement AI solutions with their own staff. There are about 40,000 active food specialty malls in Korea, indicating great growth potential for Farmkit. After launching FoodQ in the second half of last year and gaining word-of-mouth, they have already secured 30 food specialty malls as clients. Their goal is to expand this to 500 clients by next year. They also plan to enter the North American market starting with participation in CES, the world's largest electronics and IT exhibition, next year. Yang emphasized, "No customer has discontinued using Farmkit's solution yet, which proves it helps sales. We will focus on advancing the solution so that specialty malls can offer experiences comparable to large shopping malls."