Woowa Brothers unveils KAIROS rider AI, cuts dispatch time 25%

Woowa Brothers introduced KAIROS, its AI-based rider supply forecasting system, at WoowaCon 2026 in Seoul on October 28, according to techm.kr. The company said dispatch time fell by 25% after the system was introduced and that deliveries excessively concentrated on certain riders were reduced by about 50%.

KAIROS forecasts demand and supply by region and by hour, drawing on around 20 machine learning models. It is built in two parts: a prediction model layer and an automation platform that turns those predictions into supply actions with minimal human intervention.

The team defined actual rider driving time on the Baemin Connect app as supply time, and the minimum time needed to maintain delivery quality as target time. Analysis showed that 91% of regions meeting 100% or more of target time achieved delivery quality standards.

In October 2025, a data scientist at the company said the team had moved from counting how many riders are needed to measuring how many hours of riding time are needed. That person, per the same account, described a five-step system covering regional order forecasting, required supply time calculation, natural rider driving time prediction, additional supply calculation and supply target selection.