When I asked dealers what new solutions they might be looking for here at NADA, “AI” and “ChatGPT” were frequent responses. Like the rest of the world, AI is on the rise and on the minds of many.
Perhaps the most intriguing take came from a dealer who believes that AI, over time, will empower his people to make better decisions and likely lead to operating with fewer people in positions where AI can do the rote tasks individuals now perform. But the dealer shared that the AI-related he’s most interested in seeing did not appear available among the many vendors offering AI solutions here.
“I’d like to see AI tell me what vehicles will sell in two or three weeks to help my stocking strategy,” the dealer says.
The dealer’s thinking lands squarely in the four-tier framework that Cox Automotive’s chief data officer Ben Flusberg shared in our Thursday panel session at the National Auto Auction Association meeting. The breakdown:
Insights: This is where AI and machine learning can look at past data/events and provide insights on what occurred.
Predictions: This is where AI-driven tools use algorithms to predict specific outcomes, such as risk or opportunity related to the pricing or sale of a vehicle in the future.
Recommendations: Here, AI solutions recommend a particular course of action. This is the type of AI/machine learning the ProfitTime GPS performs as offers pricing and appraising recommendations on specific vehicles.
Automated decisioning: This is where an AI solution acts on the recommendation by itself, with some human oversight.
It seems to me that the automotive industry has and will continue to operate in the “predictions” and “recommendation” tiers for some time. This experience will then shape the types of automated AI-driven decisioning that take hold in dealerships and related businesses.
Here are a few other take-aways from the floor:
Profitability: Dealers are keenly aware that the new and used vehicle markets are normalizing. They see it in their financials, and the diminished front-end grosses they realize on retail sales. Not surprisingly, the dealers who complain about this inevitable shift—ours is a cyclical business, after all—in used vehicles tend to suffer from operational practices that worked pretty well in the past two or three years but work less well today. These include stocking more used vehicles than you sell in a given month, relying on the calendar to make pricing decisions and giving every vehicle the same “shot” or time in inventory to sell.
Packs: I did an informal poll of Performance Managers here at NADA. They generally agree that many dealers increased their packs on used vehicles in the past two to three years as gross profits soared. Today, the average pack seems to be around $1000, and “more dealers have been willing to go even higher.” I’ve long regarded packs as an effective tax that slows retail sales and can cause consternation among sales teams. My sense is that dealers who raised their packs would be wise to right-size them for the current market and the investment value of individual vehicles.
EVs: Almost every dealer I talked to remains wary of EVs. It’s not that they don’t recognize EVs represent the future. Rather, they understand consumer demand and related charging infrastructure isn’t as well developed as it needs to be for wide-scale adoption. It’s also clear to me that while there’s an awful lot of work underway to understand how to properly value a used EV, uncertainty abounds. It’s important that dealers and used vehicle managers recognize that the age and mileage of an EV don’t correlate to a valuation in the way they do with internal combustion vehicles. EV batteries degrade over time, even if you don’t drive the vehicle all that much. Hence, a five-year-old EV with 20,000 miles might not be worth as much as two-year-old EV with higher mileage. Still, I’d have to agree with the overall assessment that EV batteries, as a whole, are holding up better than almost everyone expected.
It’s been a long, productive day at NADA. I’m looking forward to what Day 3 will bring and I’ll share it here.
The post NADA Day 2: AI Everywhere and Some Familiar Challenges appeared first on Dale Pollak.