[SINGAPORE] Local car rental company Lumens faced difficult operational challenges when recording damages during car rental business Ryro's mobile phones.
Checking the car took an average of 10 minutes, and staff manually documented scratches and dents on the paper shape. Because manual processes tend to suffer from human error and subjective judgments, the company has collided with the idea of tapping generative artificial intelligence (GEN AI) to automate the process.
Lumens, the early initiator of the private employed vehicle industry, has been operating since 2014. With around 6,000 cars, it is now one of Singapore's largest fleet operators.
In addition to other services such as limousine services and short-term rentals, it also offers long-term rentals for private vehicle drivers.
In 2020, to increase the revenue of driver partners during Covid-19, when demand for boarding agencies fell, Lumens could set up a counted Lylo among clients' corporate organizations and call drivers for point-to-point transportation needs.
For example, after realising that visitors faced difficulties in securing rides when the Singapore Expo left the venue, Lyro worked with it to set up a streamlined process that could make it easier for visitors to book a ride.
Newsletter for you
Friday, 8:30am
sgsme
Stay up to date with Singapore's SME community with profiles, news and tips.
As lumens grow and look for areas for improving efficiency, we also look at AI solutions in areas such as predictive maintenance and personalization to each customer's preferences.
“A good example at the intersection of operational efficiency and customer experience is car checking, which is important because it has a financial impact on both the lumen and the customer.” Business time.
The automated damages reporting solution was developed last year as part of InfoCom and Media Development Authority's Gen AI X Digital Leaders initiative, helping businesses tap on GEN AI with access to resources and expertise. Lumens has an in-house technical team, but lacked the expertise to build it on its own.
The app steps guide users on how to capture photos from different angles. The AI model checks photos taken during the collection to determine if damage has occurred. If damage is detected, the system will check against the rental insurance purchased to calculate additional repair costs.
The system can also detect changes in the odometer and fuel gauge and check if the customer refills the fuel tank to the same level when returning the car. It is integrated into the billing system, which synchronizes with Lumens' fleet management system to make the entire process seamless.
This reduces the time required to perform a manual vehicle check that takes approximately 10 minutes per vehicle. The new system, in which customers are running their own car checks, can quadruple the productivity of their operational staff, allowing them to serve up to four customers simultaneously.
Sun said: “We estimate that we will save more than S$300,000 a year, and this is only in terms of operational efficiency.”
New processes increase convenience, trust and transparency that are important to customers. Lumens hopes that AI will help its customers acquire new revenue streams through this self-service process and attract a broader customer base in the growing digital mobility segment.
The development of this solution has raised challenges. One was training AI models to identify the right car model. This includes identifying the year in which it was manufactured, as the same car models have different body types over various years.
AI models also need to learn different dashboard layouts. This may also vary depending on the different production dates of the same car model. This is even more complicated. Some dashboards have analog gauges, while others are digital.
“In addition, lighting and weather conditions and photo angles also have a big impact on the results of the assessment, so the team actually put a lot of effort into model training,” Sun said.
Spoofing detection measures are also required to confirm that the photographs you took are from of? ) Same vehicle. Solving these challenges can be rewarding, Sun says, making AI solutions more robust and reliable.
Lumens can tap on the advantage of having over 10 years of car check records in paper form, along with photos of damaged vehicles to train AI models.
The solution is expected to be released towards the end of the year and is expected to help surge in demand during the holidays.
The key focus now is to increase the accuracy of more challenging models, especially under a variety of lighting conditions. Lumens is building an intuitive solution where AI guides users through the car collection and car return process through chat boxes.
All of these measures are part of a plan to improve the process that is suitable for your clients and grow your business.
