Baedal Minjok introduces AI-based delivery time prediction function

Predict time to completion of 58 types of data, including delivery status, over the past two weeks

(Seoul = Yonhap News) Reporter Hong Guk-ki = Baedal Minjok (Baemin) is introducing an artificial intelligence (AI)-based delivery time prediction function in the store delivery order acceptance process.

According to the store owner community 'Baemin Restaurant Plaza' on the 26th, Baemin will improve its system by having the business owner only input the food preparation time when accepting store delivery orders from the 1st of next month, and AI will automatically predict and calculate the delivery time.

Store delivery is not a method of direct delivery by Baedal Minjok, but an ordering method in which delivery is performed by a separate delivery agency contracted by the business owner.

Previously, business owners had to manually estimate and enter food cooking time and delivery time.

However, compared to food preparation time, delivery time fluctuates greatly depending on local congestion or delivery infrastructure conditions, making it difficult for business owners to accurately predict.

As a result, there were cases where the estimated arrival time informed to the customer was different from the actual delivery completion time.

However, with this improvement, business owners only need to enter the food preparation time (cooking time) when receiving an order, excluding the delivery time.

Delivery time is automatically predicted by AI and informed to the customer. The food preparation time that can be entered in the improved method ranges from 5 to 90 minutes.

AI predicts delivery time by combining 58 types of data, including the delivery status of each store over the past two weeks, regional congestion, delivery infrastructure status, and rider security level.

As a result of Baemin's analysis of actual store delivery order data from June 17th to 21st, the percentage of deliveries completed within the expected time range informed to customers was 30% in the existing method, but rose to 52% after applying the AI ​​prediction method.

As a result, the delivery time guidance accuracy increased by approximately 1.73 times.

A Baemin official explained, “Because business owners only need to enter the food cooking time, which is relatively easy to predict, receiving and processing store delivery orders becomes simpler,” adding, “We expect to be able to provide customers with more accurate expected arrival times.”