Got into a minor fender bender in your rental car? If the inspector doesn’t notice the damage when you return the car, you’ll dodge that bullet. But the rental-car agency will be forced to pick up the repair bill.
Considering the number of cars owned by such agencies — Enterprise Car Rentals, for example, has 1.7 million cars in operation worldwide – the cost of not identifying damage upon a car’s return can be substantial.
Netanya-based startup Click-ins has developed a technology for ensuring that damage doesn’t get overlooked.
The concept is quite simple: An inspector at a rental-car agency takes between four to eight pictures of a car before it leaves and another four to eight when it’s returned.
The pictures can be taken with a standard mobile phone and Click-ins’ software guides the inspector where and when to snap. If an image is too blurry, Click-ins will advise the inspector to move to the right or left to grab the best possible picture.
Click-ins then uses artificial intelligence, deep learning and 3D modeling to compare the pre- and post-rental pictures to each other and to the company’s massive database of car images, to accurately detect damage – and then to bill the customer.
The company’s business model is to charge a per-vehicle inspection along with a monthly subscription fee.
Click-ins, which has raised just shy of $6 million including government funding from the Israel Innovation Authority, is used by government entities in Israel.
The company has partnerships in the United States, Europe, Asia, South Africa and Russia. Strategic investor Shlomo SIXT, the Israeli licensee of the global SIXT brand of rental-car operations, is also a commercial client. Shlomo SIXT operates some 11,000 vehicles in the country and is the first to successfully implement AI-based solutions for car rental and return using Click-ins technology.
While the concept is easy to understand, the technology behind Click-ins has been a decade in the making, explains CTO Dmitry (“Dima”) Geyzersky.
Click-ins employs a hybrid of computer-managed technology and human-run inspection.
“You wouldn’t expect a robot to conduct surgery without the presence of a senior physician,” Geyzersky tells ISRAEL21c. “So, when you hand over a car to an inspector at a car-rental company, you shouldn’t rely 100 percent on the system to take pictures and assess any damage. We don’t replace a human inspector; we make his work more efficient.”
A big part of the hard work Click-ins undertook is the creation of entirely fake – but incredibly realistic-looking – images of vehicles using a technique known as “synthetic AI.”
The company’s proprietary software recreates car makes and colors, then renders them with different colors, damages and deformations under a variety of lighting conditions and backgrounds. The software automatically annotates the images, eliminating the human factor.
Why not just take photos of cars the old-fashioned way? By creating fake images, Click-ins can be ready in a jiffy when automakers release new models. It also helps avoid ethical issues when “training” the AI.
Moreover, taking pictures of actual vehicles requires hiring a team to hand annotate them. “It’s a never-ending story,” Geyzersky says.
With synthetic AI, “we don’t need to train our systems on customer data. We see it as ‘ethical AI.’ Our robots are working 24/7 creating synthetic data for our models and the AI learns from this data as if it was real.”
You can see some synthetic AI-generated images from Click-ins below. It’s nearly impossible to tell these aren’t real cars.
No hardware to buy
Click-ins employs a classic SaaS (Software-as-a-Service) business model. There’s no hardware to buy or even software to download; Click-ins runs as a mobile website on any iOS or Android device and as an API for B2B customers.
Click a link and the phone’s browser opens. When you upload images, they are automatically sent to the Click-ins cloud for analysis.
If Click-ins identifies damage, the inspector or fleet manager will get a notification highlighting visually where the damage is. Click-ins can integrate with a rental-car agency’s existing vehicle tracking software.
While most of Click-ins’ competitors are using video rather than still images, “We don’t believe in video,” Geyzersky says. “It’s less user-friendly, it takes more time, and it’s a lot of data to upload.”
Indeed, the parking garages where rental cars are returned often have limited cellular reception, making a large video upload less feasible.
The name “Click-ins” is a throwback to the company’s original focus: tackling fraud in insurance. Geyzersky’s partner, current company chairman Eugene Greenberg, worked in the insurance industry, while Geyzersky’s background is in technology and military-grade intelligence.
“When we started this journey in 2014, we began developing a huge system for fighting insurance fraud,” Geyzersky explains.
But they soon realized their most promising vertical was the image processing piece. “When we applied our solution to car rentals, it was a perfect market fit.”
Frequently, inspectors don’t mark all the damage on returned cars, so customers are not charged. This leads to millions of dollars in losses to car rental companies, Geyzersky says.
While Click-ins might seem bad for renters hoping to slip some damage past inspectors, it has a consumer benefit, too.
“Let’s say an inspector misses a few dents or scratches, so they mark two areas instead of five or six,” Geyzersky points out. “When you return the car, another inspector marks the other damages the first inspector missed and you, the most recent customer, get charged as if these are new damages – because someone didn’t do their job properly in the first place.”
Click-ins is not the only Israeli startup aiming to make life easier – and more profitable – for organizations involved in identifying vehicle damage.
Yokne’am Illit-based Spinframe Technologies provides importers, agents and automotive companies with 360-degree photo technologies for vehicle advertising, marketing and operations, including visual damage monitoring.
In Spinframe’s model, its technology can be applied on any visual device, from mobile phones to dedicated use-case adapted hardware, such as a portable “tunnel” where the image capture and comparison take place in an automated process. The company says that its product identifies damage up to 90% of the time.
Spinframe was selected out of hundreds of applicants through Konnect, the Volkswagen Group’s Open Innovation Hub in Israel, to work on applying its technology as part of group’s “service center of the future” initiative.
“Over the last three years, we have continually developed our technology with the goal of providing an innovative, versatile product that supplies advanced testing capabilities for vehicle parts while being simple to use, without requiring specialist operating skills,” said Ori Dangur, Spinframe Technologies CEO.
“This is an easy-to-integrate product that will deliver immense value to each service center in terms of assessing external vehicle damage.”
Other Spinframe partners include Avis, SEAT, Shlomo SIXT and the Ashdod Port. The company has participated in IBM Alpha Zone, NVIDIA’s Inception Program, Microsoft for Startups, Quantum Spark and 365x.