Baemin, a leading food delivery platform, has unveiled an innovative artificial intelligence (AI)-driven strategy to streamline the entire food delivery process, from order placement to final delivery. The company aims to enhance customer experience by predicting preferences and demand while seamlessly connecting restaurants, riders, and robots in real-time.
On Wednesday, Woowa Brothers, Baemin’s parent company, hosted the Woowa Tech Conference 2026 (WoowaCon) at the Grand InterContinental Seoul Parnas in Seoul’s Gangnam district.
The event, themed, Technology for a City that Never Stops, featured over 40 presentation sessions. Out of approximately 10,000 pre-registrants, about 2,000 attendees were selected through a lottery system.
During his opening keynote, Woowa Brothers Chief Executive Officer (CEO) Kim Beom-seok introduced the concept of a Seamless City as the future urban landscape Baemin envisions.
Kim explained that their goal is to create an interconnected ecosystem of customers, restaurants, riders, and robots. By leveraging demand forecasting and real-time optimization, they aim to minimize unnecessary wait times and ensure customers receive their orders precisely when needed.
The waiting begins even before the customer hits the order button, Kim stated. It’s building a city where everything is prepared in advance, constantly connected, and moving to anticipate and meet customer needs.
AI Advancements: From Customer Preferences to Cooking Times
Woowa Brothers’ Chief Technology Officer (CTO) Ko Myung-seok showcased their major AI-driven technological achievements.
Baemin has developed a sophisticated hyper-personalization engine that utilizes over 3,000 customer-related variables to deliver tailored services.
The AI agent determines the most suitable benefits and recommendations for each customer, continuously improving its prediction model based on user interactions and outcomes.
The company has also made significant strides in cooking time prediction technology. Moving beyond the traditional method of relying on restaurant owners’ manual input, they’ve developed an advanced model that recommends optimal cooking times. This model factors in order volume and processing speed for each menu item. According to the company, this innovation has slashed the average error rate in cooking time predictions by 50%.
For delivery time predictions, Baemin’s system analyzes 58 distinct data points, including store-specific delivery status, regional traffic congestion, and delivery infrastructure. Once restaurant owners input food preparation times, the AI calculates the expected delivery duration. The company reports that this approach has more than doubled the accuracy of delivery time predictions compared to previous methods.
Ko emphasized that it’s developing technology that not only predicts and makes judgments more accurately but also executes autonomously and refines its predictions based on real-world results.
Anticipating Rider Demand Before Rainfall: Response Time Cut by 80%
Daniel Han, head of Delivery, illustrated the practical application of their AI technology using the example of managing orders during rainy weekday evenings.
In the past, the operations team had to manually adjust delivery zones by region after rainfall began. Now, their system proactively forecasts rider demand during peak hours for each area, allowing them to secure necessary personnel in advance.
This proactive approach has reportedly slashed rider demand response times during inclement weather by approximately 80%.
The company is also optimizing rider routes using real-world driving data. Recognizing that the shortest path on a map doesn’t always translate to the quickest route due to factors like hills, intersections, and road conditions, they’re developing more efficient routes by incorporating actual driving data and experimental results.
Baemin has also introduced an automatic recovery system for order processing errors. Previously, when service disruptions or message failures halted order processing, human intervention was necessary. The new recovery system allows operations to resume from intermediate stages automatically.
In select areas, the company has launched a service where robots deliver B-Mart orders directly to customers’ doorsteps. These robots reportedly operate autonomously for over 99% of their total travel distance, covering approximately 5,000 km (about 3,106 miles) monthly.
During the event, Woowa Brothers showcased a wide array of technological innovations, spanning AI, machine learning, backend and frontend development, mobile apps, data analytics, and robotics. The conference also featured mentoring sessions and hands-on programs, fostering interaction between current employees and attendees.
