BGF Retail, which operates the convenience store CU, decided to ask "fake customers" created by artificial intelligence (AI) about reactions to new products instead of real consumers. By predicting consumer reactions with AI without conducting surveys of people, the company aims to cut the time and expense required to plan new products such as lunch boxes and desserts.
BGF Retail said on the 15th that it signed a "strategic business agreement (MOU) for data-driven retail innovation" with IntelliCIA, an AI synthetic consumer technology startup. The signing ceremony held on the 14th at BGF Retail's headquarters in Gangnam-gu, Seoul, was attended by Lee Eun-gwan, BGF Retail CX division head, and Baek Seung-guk, IntelliCIA CEO, among others.
A "synthetic consumer" is a technology that predicts reactions to new products, prices, and promotions by having AI create virtual consumers on the scale of millions without real consumer research. Within a large language model (LLM), it generates thousands to tens of thousands of virtual personas with varying ages, genders, and preferences, and has them respond to surveys like real people. Before launching a new lunch box, for example, AI first simulates purchase intent and preferred price ranges.
BGF Retail's stated strengths are speed and expense. Traditional consumer research involves conducting surveys and interviews with thousands of people, taking months to produce results, and often only evaluating products after launch. In contrast, using synthetic consumers can shorten the research period to three to five days and significantly lower the expense, while enabling repeated validation whenever needed, the company said.
The two companies have already been conducting a proof of concept (PoC) on convenience store lunch boxes since May. In traditional research, responses confirmed that consumers wanted "healthy lunch boxes," but it was difficult to consolidate the complex demand for "healthy yet tasty" into actual product planning. The company said the synthetic consumer analyzed criteria for balancing taste and health and turned this vague dual need into concrete product insights.
BGF Retail will also use IntelliCIA's store digital twin solution ParaStore. The plan is to replicate actual CU stores in virtual space and simulate product displays, customer traffic flows, and merchandise composition (MD) in advance to reduce trial and error. Lee Eun-gwan, the division head, said, "CU will more precisely reflect changing customer needs and create a new retail experience that provides personalized products and services."