Photo/Illutration A Keio Liner reserved-seat train (Provided by Keio Corp.)

Gone are the days when a laborious search by hand is required to locate a missing item at the lost and found office.

In May, Tokyo-based private railroad operator Keio Corp. introduced a service that relies on artificial intelligence (AI) to sort things out quickly.

Kyushu Railway Co. (JR Kyushu), based in Fukuoka, followed suit by starting to provide a similar service in August.

Claimants simply need a smartphone to make an inquiry.

First, they need to enter a description of the lost article, including any defining features, along with a photo, and date and place of loss, via an official account of the lost property search service on the Line free messaging app.

The AI system taps into a database to compare the inputted information and selects candidate articles that best match the description. Once located, the lost article will be returned to the claimant at a station or elsewhere after proof of identity is established.

Find Inc., a Tokyo-based startup firm, developed the lost property search system.

A record 4.15 million lost objects were picked up in Tokyo in 2019, according to the website of the capital's Metropolitan Police Department.

The figure dropped to 2.81 million in 2020 partly because much fewer people ventured outdoors while the novel coronavirus was raging. However, the figure rebounded to 3.43 million in 2022.

Managing lost articles and responding to inquiries from claimants imposes a heavy burden on railroad operators, already reeling from a sharp drop in earnings due to the pandemic.

Many railway companies receive inquiries from claimants by phone or in person at their stations and lost and found offices. Checking and comparing the descriptions alone is a hugely time-consuming process.

The upshot is that AI has sharply reduced the workload of the railway operators.

Staff at lost and found offices snap a photo of each lost article that comes in, typically by smartphone for official use, and registers the information on a database. The articles are then stored with a sticky label carrying a Quick Response (QR) code that provides a link to the registered description.

The rest is left to the AI system, which uses search parameters based on the information provided by the claimant.

The rate of return of lost articles has approximately tripled since the AI system was introduced, according to Keio officials, who said the company handles around 16,000 lost items each month.

“That’s presumably because the service allows inquiries to be made conveniently on a smartphone,” a Keio official said.

“The system is reducing the workload of station staff as well,” a JR Kyushu official noted.

Related News