AI Rental Scanner Unfairly Charging You With Damage? There's AI For That

The Drive

3 days ago

The latest car news, reviews, and features.
The way things are going, it's very reasonable to expect that the next time you rent a car, a human won't be reviewing its condition when you bring it back. Hertz and Sixt are bringing AI-powered vehicle inspection scanners to more of their lots, with competitors taking notice. And, if you decide to contest one of their accusations of damage, you may be using AI for that, too.
Since publishing our initial stories on this trend, a few digital inspection startups have reached out to share how they're leveraging the tech, but not quite in the way that UVeye and ProovStation, the companies that are partnering with Hertz and Sixt, respectively, are handling things. Using only your smartphone, this software allows you to capture photos and videos to create your own digitally verified record of a car's state at pickup and delivery. Then, if the rental agency gets on your case about an alleged dent that is clearly a speck of dirt, you've got the evidence to back you up—provided the customer service rep listens, of course.
Interestingly, one of these startups—Ravin AI—actually has some familiarity with the rental industry, because it worked with Avis in 2019 and Hertz in 2022 as those companies experimented with AI inspections for the first time. These days, Ravin isn't focusing on the rental side of things, electing instead to work with insurance companies and dealerships. It's partnered with IAG, Australia and New Zealand's largest insurance firm, to allow customers to scan their vehicles and receive quick quotes on damage repairs using their phones. ProovStation says that its rental scanners can be installed anywhere in three hours and capture about 250 pictures at 4K resolution to spot damage. ProovStation
Ravin's decision to focus on insurance, rather than rentals, can at least in part be attributed to the very horror stories we've been hearing about in recent weeks, founder and CEO Eliron Ekstein told me.
'When you work for a car rental company, if you go about this in the wrong way, then you're actually going against their interests. Because you're going against their customers in the end,' Ekstein told me. 'And we quickly realized that if we maximize our proceeds in that business, we're actually going against their customers and themselves at the end of the day.'
Ekstein believes that the potential still exists for rental agencies to use AI in a way that creates efficiency in their systems without causing strife for consumers, but right now, they're attaching exorbitant fees to every insignificant blemish likely because 'they need to justify the cost' of their scanning equipment. And the companies that are developing this tech aren't exactly shy about the potential. ProovStation's About Us page says, right at the top, that the company's 'cutting-edge AI-powered technology [empowers] businesses by turning routine inspections into gold mines of untapped opportunities.'
'You need to be super transparent with your customer,' Ekstein said. 'Because those that really damage the car will eventually admit it, if [the rental company has] evidence, but they need to damage the car in such a way there is actually economic impact. You can't go after scuffs—that goes against the whole service culture of those companies. I just won't rent from you again. But, if I damage the door in such a way that's going to cost you at least $700 or so to repair it, then it's fair for you to go after me. And I'll also want to see the estimate behind that—I'm not going to pay you just a flat fee.'
Ekstein says Ravin has poured 10 years of data and two billion images into training the company's system to properly assess costs for exterior damage. He told me that these recent reports of poor customer experiences with rental scanners prompted him to open up Ravin's tech to the public, so that anyone can use it on their smartphone for free. In the bottom image, you can see how Ravin's scanning process works. You frame your car in the outline, which turns green when the system can appropriately capture the car. After doing this several times around the whole vehicle, it'll call out areas where it notices damage, and ask you to take a closer look if necessary. It'll even guide you to the spot using onscreen arrows. Adam Ismail
I took it for a spin myself (it's accessible through the 'demo' link on Ravin's homepage) and it was easy to use, though not entirely accurate. The system failed to call out two obvious chips on my car's driver's side front fender, but misidentified a reflection in the hood's stamping as 'minor damage.' My car was admittedly quite dirty when I tried it, which probably didn't help things—but rentals surely won't be showroom quality, either. Conditions and lighting during photo capture are actually one of the biggest challenges in this space, Ekstein told me, as 'who takes the image, what time, [and] what angle' can produce dramatically different results.
When all is said and done, Ravin sends you a report to outline the differences between your before and after photos, which you can share as you please. And Ravin isn't the only one; there's another startup named Proofr with a similar purpose, though it's advertised for assessing rental properties as well as cars. I tried out Proofr's app as well, and while the interface is very slick, it did repeatedly crash on me while saving images. Unlike Ravin, Proofr runs $9.90 a month, with a three-day free trial for new users.
'More competition is also highlighting what's important and getting customer feedback, and whoever has the best tool will win,' Ekstein said, summing up this burgeoning industry. I hope he's right, because history has more than its share of inferior tech formats and systems winning out. Right now, the bar for AI scanners is so low that we've first got to overcome the hurdle of telling apart rock chips from dirt. A long way to go, indeed.
Got a tip? Let us know at [e-mail]