Sourcing used inventory at the right price and accurately assessing reconditioning costs are two pain points for dealerships. In days past, dealers often had to rely on their gut to handle the two tasks.

No more. Cox Automotive’s vAuto suite of product uses artificial intelligence to help dealers buy the best used inventory mix including accurately estimating reconditioning costs to retain desired profit margins.

“You make your money when you buy a car,” Derek Hansen, Cox Automotive’s senior vice president of dealer, lender and inventory management solutions, told Auto Remarketing. “Using the AI tools to really solidify the price you are paying at the front end is where the opportunity is.”

Finding recon gaps

Among the tools: An AI-enabled enhancement that is using data to assess how a dealership’s team performs across the different acquisition channels, including auction, trade-ins, service drive and private party, Hansen said.

vAuto’s Global Search uses AI to track performance by appraiser and by channel and provides more data so appraisers can more accurately estimate recon costs.
Poorly estimating recon costs can cause dealers to miss out on buying the most profitable used inventory, Hansen said.

For example, a dealership estimates a $1,500 recon cost for a vehicle acquired at auction. If it is in worse condition and costs more to recon, that eats into the vehicle’s profitability.

But what if it needs less reconditioning than the estimate? A dealer may not have purchased that used vehicle due to an overly aggressive recon cost estimate and so missed out on a good opportunity, Hansen said.

AI can also help dealers more profitably acquire off-brand trade-in or service drive used inventory, Hansen said.

For example, a 2019 Subaru pulls into a Ford dealer’s service drive. AI can tell the dealer what primary maintenance issues to look for in that Subaru by looking at every resource on the internet, including Reddit forums, OEM manuals and Kelley Blue Book write ups.

Some help from an AI assistant

Did you ever wish you had an assistant that could handle all your inventory management tasks, streamline workflows and make actionable recommendations? Well, wish no more.

Profit Time Assistant, an agentic AI tool within vAuto, is “starting to pick up” with dealers, Hansen said.

Some examples of Profit Time Assistant’s capabilities are: It would identify if a car was missing an odometer reading on the listing and recommend the odometer reading be added. Or, it would notice which vehicles had not had a price adjustment in a week and recommend new prices for them, Hansen said.

Importantly, Profit Time Assistant just identifies areas where action is called for and makes a recommendation. An actual human must accept the action, he said.

That helps avoid AI hallucinations which are “a very real problem,” Hansen said. “AI at its core is trying to solve a problem. It’s trying to solve a prompt. And when it can’t find that solution, it’s very rare that it comes back and says, well, I couldn’t find an answer. It’s going to just tell you what it thinks it should, right?” he explained.

Cox has very rigorous data cleaning, aggregation and testing protocols to ensure AI-generated actions are reliable, and human oversight is also integrated into the workflow as another safety measure, Hansen stressed.

Lean into AI

Every two years, Cox Automotive holds a vAuto client meeting to share advances and get feedback.

The meeting was held this year in Austin, Texas, March 31 through April 1. Seventy five of its “most progressive clients” representing some 700 rooftops were flown in to participate, Hansen said.

After walking them through the AI features Cox released over the last year and discussing where Cox is heading for the next 12 months, the dealers were broken into smaller groups where they could role play, interacting with the AI tools and asking questions, Hansen said.

The top takeaway from the meeting regarding AI in the dealership, he said, is “if you’re not embracing it, you need to lean in.”

The other main takeaway? Dealers need to be testing tools now because AI is moving so fast, he said.

“Because of that speed, again, continuing to test, continuing to lean in, continuing to have your operations exposed to where there are specific problems or jobs to be done that AI can be deployed, that can help you make better decisions that can help you make faster decisions, that can help you save money,” Hansen said.