CCC on Wednesday announced the availability of a damage detecting artificial intelligence which can scrutinize any body type based on images alone.
It also showed off how the system generates “heat maps” which can highlight that damage to customers — or adjusters.
“Research shows that 9 out of 10 consumers prefer estimates when delivered with heat maps that clearly highlight the damage versus estimates delivered without heat maps,” the information provider wrote in a news release.
The AI photo damage detection capability announced Wednesday is different from AI photo damage analysis, CCC product management Vice President David Bliss explained. However, the two can work together; Bliss said the detection technology powers CCC’s “Smart Total Loss” photo analysis product unveiled last year.
He said CCC had through machine learning trained the artificial intelligence to recognize vehicle damage without having to rely on a wireframe overlay of the vehicle class for reference. (For example, comparing a Ram 1500 distorted by a collision to a shape of an undamaged pickup.)
“That was a huge leap forward in the technology,” Bliss said. The new deep learning approach made the system smart and flexible enough that it could be applied to “virtually any car” based solely off photos; there was no need for CCC to find a 3-D model of the vehicle first.
Bliss said the technology is able to handle close-ups of minor damages like nicks and scratches as well as more traditional vehicle poses. “It is very accurate,” he said.
However, one can see applications besides impressing customers and providing a hedge against an appraiser missing damage in the initial inspection. We suggested shops could use it as a higher-tech version of the damage-highlighting Dent Viewer grids, proving to a remote adjuster that an impact is truly as large as the shop insists — the AI and heat map says so.
“Your use case is absolutely correct,” Bliss said.
Some competitors are also applying damage-detecting AI to count and classify hail dings in a vehicle. Bliss said that for now, CCC’s system works for collision damage — but “we are working on that concept” as well.
“It’s a little more difficult with hail,” Bliss said, for a vehicle owner or other photographer would need to produce images of the roof as well.
The system can work indoors or outdoors, day or night, and an associated app can trigger the flash when necessary, according to Bliss. The app can scan the vehicle VIN and direct the user to the appropriate photo angles for that vehicle class, he said.
The system does have limitations. It can’t detect hidden damage, and it can’t deduce the entire area that would need to be manipulated with hammer and dolly during a traditional dent repair.
Bliss said commercial availability of the damage detection system was “not far at all.” It would be incorporated into insurer-facing products this month and by the end of May, he said.
As noted above, scenarios exist where a shop might wish to use such a tool to wow customers themselves or provide documentation to an insurer.
“The technology could certainly be used by shops,” Bliss said, and he said his goal was to have it available for collision repairers in CCC ONE and its associated products, such as CCC ONE Touch. However, that rollout will be “not that soon” for collision repairers as the , Bliss said, reasoning that such software wasn’t as critical for a party who had the vehicle available to inspect in person.
“For photos to deliver real benefits they need to be supported by data and a deep understanding of how the claims process works,” Bliss said in a statement. “We’re injecting our claims solution suite with AI, which becomes infinitely more powerful when combined with data in CCC ONE, and CCC’s 30 years of experience supporting the auto claims industry. Our damage detection technology and heat maps bring depth and insight into otherwise opaque damage photos. And we know through our 50 million annual consumer touchpoints that transparency is linked to satisfaction. CCC’s real-world applications of AI are a gamechanger for the industry.”
Bliss said CCC has talked to shop about the damage detection tools and will have a test version; contact your CCC account rep if you’d like to participate.
CCC, April 25, 2018
CCC’s AI damage detection system examined the damage on this Getty Images photo and represented what it detected as a heat map. (Provided by CCC/Getty Images)
CCC’s AI damage detection system examined the damage on this Getty Images photo and represented what it detected as a heat map. (Provided by CCC)