Baemin Unveils AI-Powered Seamless City Vision

Baedal Minjok (Baemin) presented “Seamless City” as a vision for the future at Woowa Brothers’ Woowa Tech Conference 2026, held on the 28th at Grand InterContinental Seoul Parnas in Gangnam, Seoul. The concept aims to predict the food and goods customers need and connect customers, stores, couriers and robots to reduce waiting. Woowa Brothers CEO Kim Beom-seok said AI would understand customer intent and, within the scope customers permit, automatically handle product searches and selections as well as repeat orders.

Kim identified prediction, connection and optimization as the main implementation challenges. He said the system must forecast orders before they occur, connect customers, stores, couriers and robots in one system, and optimize multiple decisions amid sudden rain or traffic congestion. Baemin applied a “hyper-personalization engine” that uses more than 3,000 customer variables and an ontology (a structured map of concepts and their relationships) to help AI agents assess benefits and recommendations. A cooking-time prediction model that accounts for order volumes and processing speeds by menu item cut average error by 50%. Delivery-time predictions using 58 data points, including local congestion, more than doubled in accuracy.

Woowa Brothers Chief Technology Officer (CTO) Ko Myeong-seok called the system connecting customers, stores and couriers a “Real-time Physical Network.” He said the company is developing technology that acts on predictions and decisions, then improves predictions based on the results. In July, Baemin introduced Kairos, an automated platform for managing courier supply. Previously, operations staff made those decisions directly; Kairos uses AI models to calculate how many couriers are needed and when. Baemin said it operates more than 20 machine-learning models for accurate forecasts and has cut response time to rainy-weather supply shortages by about 80% by predicting them in advance.

Baemin is also applying AI to arrival-time predictions, real-time cooking-time estimates for store owners and route optimization. Baemin B Mart delivery robots travel up to 5,000 km per month in areas including Gangnam, Seoul, and complete more than 99% of actual journeys autonomously. The company built an internal Model Context Protocol Hub (MCP HUB) that provides AI agents, and also developed the AI agent “Experiment Dog” and the internal web tool “PA Admin.” Its “AI Search,” which finds food and products through conversation, has been introduced as a pilot.

WoowaCon, which Baemin has held since 2020, took place this year under the theme “Technology for a City That Never Stops” and featured about 40 presentation sessions. There were 10,000 pre-registrants, and 2,000 people selected by lottery attended. The event also included mentoring, the mobile mini-game “B Mart Tycoon” and “Prompt Speedrun,” a contest to see who could write prompts fastest while viewing a screen.

Baedal Minjok (Baemin) presented “Seamless City” as a vision for the future at Woowa Brothers’ Woowa Tech Conference 2026, held on the 28th at Grand InterContinental Seoul Parnas in Gangnam, Seoul. The concept aims to predict the food and goods customers need and connect customers, stores, couriers and robots to reduce waiting. Woowa Brothers CEO Kim Beom-seok said AI would understand customer intent and, within the scope customers permit, automatically handle product searches and selections as well as repeat orders.

Kim identified prediction, connection and optimization as the main implementation challenges. He said the system must forecast orders before they occur, connect customers, stores, couriers and robots in one system, and optimize multiple decisions amid sudden rain or traffic congestion. Baemin applied a “hyper-personalization engine” that uses more than 3,000 customer variables and an ontology (a structured map of concepts and their relationships) to help AI agents assess benefits and recommendations. A cooking-time prediction model that accounts for order volumes and processing speeds by menu item cut average error by 50%. Delivery-time predictions using 58 data points, including local congestion, more than doubled in accuracy.

Woowa Brothers Chief Technology Officer (CTO) Ko Myeong-seok called the system connecting customers, stores and couriers a “Real-time Physical Network.” He said the company is developing technology that acts on predictions and decisions, then improves predictions based on the results. In July, Baemin introduced Kairos, an automated platform for managing courier supply. Previously, operations staff made those decisions directly; Kairos uses AI models to calculate how many couriers are needed and when. Baemin said it operates more than 20 machine-learning models for accurate forecasts and has cut response time to rainy-weather supply shortages by about 80% by predicting them in advance.

Baemin is also applying AI to arrival-time predictions, real-time cooking-time estimates for store owners and route optimization. Baemin B Mart delivery robots travel up to 5,000 km per month in areas including Gangnam, Seoul, and complete more than 99% of actual journeys autonomously. The company built an internal Model Context Protocol Hub (MCP HUB) that provides AI agents, and also developed the AI agent “Experiment Dog” and the internal web tool “PA Admin.” Its “AI Search,” which finds food and products through conversation, has been introduced as a pilot.

WoowaCon, which Baemin has held since 2020, took place this year under the theme “Technology for a City That Never Stops” and featured about 40 presentation sessions. There were 10,000 pre-registrants, and 2,000 people selected by lottery attended. The event also included mentoring, the mobile mini-game “B Mart Tycoon” and “Prompt Speedrun,” a contest to see who could write prompts fastest while viewing a screen.