Toronto, Ontario — This past Tuesday, the International Roundtable on Auto Recycling (IRT) hosted its webinar, “How Data and Designation Are Transforming Auto Recycling,” featuring industry leaders Jeff Schroder from Car-Part.com and Dave Picking from Eladene Systems.
The session focused on the pivotal role of data integration and artificial intelligence in reshaping the auto recycling landscape, enabling repairers to refine their search for parts and for auto recyclers to enhance operational efficiency.
If you missed it, check out our highlights below.
Want to succeed? Get down with the data
Schroder explained that the repairer mindset often rests in the search for “the best part.” With Car-Part.com’s marketplace, users decide their own criteria for the best possible part—but the results don’t end there.
“Many will say the ‘best part’ is the part that can be delivered today. But the Car-Part.com marketplace allows users to customize their parts search journey based on expandable criteria the database provides.”
The marketplace essentially focuses on the parameters provided in the search for initial results, but then Car-Part.com leverages its internal data to provide other suggested parts to repairers, beyond their original search criteria. After all, the best part also needs to be the right part.
“You’ll often find repairers will find a better part than what they may have initially searched for, even if it’s a part that’s arriving tomorrow or the next day.”
In short: the customer defines the best part, where the Car-Part.com system defines the right part based on the original parameters.
Leveraging this data integration is a two-way street. Car-Part.com needs to work with the correct information to create the best end-user experience. Therefore, the recycler must provide the correct information from the get-go.
“Systems need to be able to answer interchange questions to source the right part. We have to rely on good data to ensure ARA Part Grading,” said Schroder. “With bad data, the search process essentially breaks down.”
The system also takes accurate delivery of parts and parts availability into account during Car-Part.com’s search process.
“We at Car-Part.com do a lot of work to ensure we have quality data to confirm the part does hold an ARA Part Grade. We have to do a fair amount of work to ensure we have clean data.”
To do this, Car-Part.com extends its services to provide audits in its management systems.
“We can show recyclers where their problems are. The auditing capabilities can show discrepancies between our system and what the recycler puts in there to ensure great data.
“We do what we can—provide the proper algorithms and such—but the ultimate source of truth is the recycler.”
The seemingly endless benefits of AI
Eladene is a U.K.-developed parts sourcing program that leverages AI in more ways than one. Dave Picking, director of Eladene Systems, started developing the program in 2013 after finding the industry frustrated with the lack of “platform-based” software.
He explained there are quite a few different version of AI that can be employed in the recycling space. Firstly, machine learning.
“That’s the side with tools like photo background removal, text-to-speech, image-to-text, et cetera,” said Picking.
Eladene uses machine learning in its Image Enhancement Model, which enhances pre-existing images of car parts.
“We wanted images to get an uplift when they came through. With Google and eBay algorithms, the highest score you can get usually involves a photo with a white background. Our tool ensures all images are uniform, and this has resulted in higher prices and lower return rates for our yards.”
Eladene’s text extraction and verification model can also identify parts based on labels. The AI looks at the part, said Picking, and knows what is on the label. It can then produce the OEM number and other valuable information.
“It’s very difficult to train the model on certain parts, but it’s a developing tool.”
Other AI-powered features Eladene uses includes parts identification and sourcing, so the system may scan what parts a car has before it arrives to the yard; demand prediction; fitment verification; operational efficiency enhancements in shipping processes, inventory management and more; as well as predicting of recycling targets to better understand demand for parts and commodity basis.
Perhaps the most exciting feature of the future, in Picking’s opinion, is the potential for deep learning AI to enhance the customer experience process.
“We’ve successfully trained an AI customer experience representative in a particular rep’s accent in under 15 minutes,” he explained. “Which in turn makes the customer feel more comfortable buying from a local dialect, as opposed to someone in England speaking with an American voice or otherwise.
“The AI is just getting cleverer and cleverer,” he added. “That really excites me—to see where customer service could soon be in our industry.”
When it comes to Car-Part.com’s AI integrations, Schroder said he’s excited for the future based on what the company is currently working on.
“We use Tableau—a data analytics tool that analyzes marketplace data. None of their original tool was based on AI learning; it’s based on statistical analysis and it essentially scans data and explains exactly what you’re looking at.
“They recently released their AI tool, which is particularly interesting as it includes the program’s already powerful algorithms. What that means is the AI doesn’t have to start from scratch—it’s layered on top of the algorithms we already use.
“Instead of using computer vision to figure out if a part fits, we use historical data analysis to figure out if the part fits. That’s the journey we’re on right now.”
Schroder emphasized that estimating systems that employ AI—particularly in the collision repair space—often automatically source OEM, aftermarket, remanufactured and OEM discount parts. Recycled parts typically pop up last.
“We need to take these AI parameters into account to stay competitive. People want the easy button. Our automatic parts sourcing helps avoid the bias of searching for a recycled part,” he explained. “If you can [search for parts] automatically and do it well, you are evening the AI-influenced playing field.”
“If you, the recycler; the marketplace and the AI engines take that into account, that’s an advantage to our industry,” concluded Schroder.
