In September 2023, I wrote a LinkedIn article titled “How Artificial Intelligence Could Improve Vehicle Remarketing & Evaluations & Reduce Overall Portfolio Risks” (click here to read it). I’m revisiting parts of it here because the ideas are even more relevant today than when I wrote them, even if we aren’t as close to that reality as I’d hoped we would be by now. If I’ve learned one thing, it’s that technology takes time to evolve before it lands front and center. In 2026, AI certainly has.

Better data standards in a fragmented industry

My proposition in 2023 was this: what if sales data were aggregated, either through industry collaboration or through AI models, so sellers could make better, faster decisions about where each specific vehicle should be sold? Imagine a model that weighs all of these at once:

  • recent marketplace performance for that year, make, model, and mileage
  • days’ supply matched to geographic retail demand through dealer inventory tools like vAuto and ACV MAX
  • damage assessments and which reconditioning truly adds value
  • vehicle history and title brand impacts
  • regional factors like seasonality and the best time for fleets to cycle units
  • transportation costs to get the vehicle to where it will bring top dollar

We can do each of these individually today. AI could truly change the remarketing value proposition by bringing them together, but only if it’s fed the right data. That’s especially true for those of us with a lot of variety in our portfolios. We aren’t like OEMs, which sell a narrower set of models. I started building a tool with some sample data:

Article content
CarGuy4U Remarketing Channel Optimizer Tool

The potential doesn’t stop at the point of sale. Consider how we evaluate portfolios and set residuals. Most of us pull historical data from our own portfolios, weigh market data and expert residual forecasts, and factor in future risks. AI could combine those sources to speed up evaluations and predict residuals more precisely.

It could account for the volumes others are bringing to market, how new models and technology shifts will affect values, and how broader economic scenarios might play out. We all do some version of this today, but it’s in-depth, time-consuming work. AI won’t predict the future, but it can give us the insight to forecast it more accurately.

Why has AI adoption been slow in remarketing?

Since the pandemic, and honestly in the years leading up to it, strong used-car demand and healthy lease residuals have bred a general complacency in remarketing. There are exceptions, but a function that once drew constant attention from analysts and banking executives has become an also-ran in many organizations.

It’s been a while since executives called the head of remarketing every week asking about returns. I remember those days well, and I got used to having the data to back up every claim.

Yes, during the pandemic you could have sold vehicles at the end of your cul-de-sac to dealers desperate for used inventory. That cushion won’t last forever. We’re already seeing erosion in subprime and repossessions.

Off-lease volumes are also rising, particularly EVs carrying high residuals, so the need to strengthen remarketing strategies will become much more apparent. I don’t expect a downturn as painful as the late 2000s were for retentions. But once a few fires get lit, you can bet leaders will start looking hard at the opportunities and technologies that have evolved over the last five or six years.

Case in point

When you keep doing what you’ve always done but the environment changes, you shouldn’t expect the same results. In 2006–2008, I was at a large bank with a heavy concentration of a make that wasn’t easy to sell. Taking those cars to the same auctions and getting slaughtered was not an acceptable answer to mounting losses. That’s when my team and I knew we had to find alternative channels. Would we have done it if losses weren’t climbing? Maybe not.

But the discomfort pushed us, and the bank, to challenge the status quo. We tried online sales from grounding locations and stepped up pre-term sale negotiations with our lease customers. As I mentioned earlier in this series, it was a game-changer for us and for many others in the industry.

Can AI fix the condition report problem?

Can AI help with condition reports, both for off-lease inspections and for upstream sales? Maybe. So far, self-inspections for off-lease vehicles have been underwhelming. Experience has taught us that customers aren’t always as forthcoming as advocates predicted. But AI damage detection is already in use in Europe.

Pairing it with self-inspections could move the needle in the U.S. by catching undisclosed damage and measuring the key off-lease damage areas more accurately. The timing is right, with a wave of off-lease vehicles on the near horizon. For now, I’m more bullish on this technology for the customer turn-in side than for downstream selling. The open question is whether it’s ready for showtime, and how much human touch it will still need to support resale and keep arbitrations in check.

Having worked the wholesale acquisition side, I know how quickly dealers will find a valid reason not to buy your vehicles. It happens when the merchandise isn’t as described, or when it has problems that a trained inspector would (or should) have caught but an AI tool may routinely miss or simply isn’t built to detect.

People aren’t perfect either. But inspectors with the right tools can sometimes identify issues that a photo-based walkaround can’t. The goal shouldn’t be to match what we do today. It should be a better experience for the off-lease customer and the leasing bank, and greater dealer confidence in downstream offerings. Getting there will take some massaging. Flexible buyer arbitration policies on these vehicles can help build that confidence along the way.

Should we be scared of AI in remarketing?

No, but we should be engaged. Used vehicles vary widely, and so do the economic forces acting on the car market. Used-car sales are never static; they require constant monitoring and adjustment. Whatever opportunities AI gives us will need human oversight to refine them and track their accuracy. And the data models behind AI will be built and maintained by warm-blooded, breathing people.

I got a taste of this evaluating AI models being built for remarketing. The strengths were real. The models did a great job of aggregating data and spotting trends, often small ones that a person scanning reports might miss but that still mattered.

Where they fell short was context. I often found they weren’t fully accounting for seasonality, regional demand, or concentration concerns, three factors anyone who has moved vehicles across the country knows can swing values significantly. That’s exactly why the human element isn’t going away. AI is excellent at telling you what the data says. It still takes experience to know what the market is doing and how it impacts the data.

Redefining auction borders

In my second article in this series, I asked whether the industry would embrace selling without the physical lane auction we’ve relied on for decades. That question applies to sellers and, just as importantly, to dealers. So far, marshaling facilities with reconditioning options have been fairly isolated.

I expect that to grow with Copart’s pending acquisition of ACV and the competitive response it’s likely to spark. (In the interest of full disclosure, I’m a former ACV team member, and the views here are my own.) The key will be whether large sellers and dealers are willing to change their habits and rely even more on digital channels and good condition reports. These new offerings may not have a physical lane or be as inviting to on-site buyers as today’s auctions. How hybrid they are may determine their success and adoption across the market.

Buyers usually follow the right seller offerings, especially if used inventory stays tight. But real change requires convincing both sides of the table. Focus on only one, and success will be limited. Cars without committed buyers leave too much retention exposure, because sellers won’t want to move vehicles multiple times without concessions. And plenty of buyers without the cars won’t last.

Looking ahead

I’ll freely admit I’m dreaming a bit here, probably thinking beyond the current limits of technology and data sharing. But with the right push from sellers, buyers, and the remarketing support community, I believe these advancements can become reality. I learned early not to fear change, especially technological change.

Everything needs controls and monitoring. Still, when I think about how far we’ve come in my 25 years in this business, and how much faster technology is now advancing in areas that could benefit this industry, I’m pumped for what’s next.

At the end of the day, though, remarketing is still a people and relationship business. We’re fortunate that this industry is full of intelligent, caring people who are committed to its continued success.

This is the 4th and final part of the series, “Both Sides of the Gate: 25 years of remarketing, from the seller’s desk and the auction block.”

It was originally posted here: Part 4: What the future may hold: AI in remarketing, new technology, redefining auction borders, and improved analytics. | LinkedIn

About the author:

I’m a 25+ year veteran of the remarketing and automotive finance world, and an avid car enthusiast. Over that career I’ve managed bank and fleet lease-end operations, 1st- and 3rd-party remarketing, residual value setting and risk mitigation, and the sale of well over 1 million off-lease and repossessed vehicles. I enjoy sharing that experience with others in the industry — whether they’re new to automotive finance or seasoned veterans — as well as helping everyday consumers buy or sell vehicles with more confidence.

I’m actively seeking full-time opportunities in this space, and I’m also open to consulting engagements, including remarketing portfolio and strategy reviews. Reach out — I’d welcome the conversation.