What Hertz’s recent AI-Scanner Fiasco Tells Us About Value-Based Pricing
Just because you can extract economic value from the individual customer using AI doesn’t mean it’s a smart or sustainable idea.
This story began about three months ago, in mid-April 2025, on automotive sites like Automotive News and Motor1. This is when Hertz announced that it was introducing automated vehicle inspection systems at its locations at major US airports. The technology was described as “advanced scanning technology dubbed an "MRI for Vehicles" to capture full images of the interior, undercarriage, and tires. It even provides a 360-degree scan that will detect chips, cracks, and paint issues.”
In a press release, Hertz explained the reasoning behind introducing this technology this way:
“With more than half a million vehicles around the globe, keeping vehicles in well maintained condition before, during, and after rentals is a critical priority for Hertz. Vehicle assessments in the rental industry have always relied heavily on manual inspections conducted in varying conditions. By implementing UVeye’s advanced AI-driven inspection technology, Hertz can significantly enhance the frequency, accuracy, and efficiency of its vehicle maintenance processes, ensuring reliable service, improved vehicle availability and transparency for its customers… UVeye’s AI-powered camera systems and machine learning algorithms enable real-time, automated inspections of a vehicle’s body, glass, tires, and undercarriage. The technology enables improvements in safety and vehicle availability by detecting maintenance issues with unprecedented speed and precision. UVeye’s tire treadwear system captures high-resolution images that are instantly analyzed to determine whether a tire needs replacement, reducing the need for manual checks and ensuring timely, proactive service. By complementing manual checks with UVeye’s technology, customers will enjoy more efficient and transparent automated checks when picking up and dropping off their vehicles.”
Pricing the damage sustained during the car rental period
Although Hertz talked solely about maintenance and customer benefits1, consumer advocates quickly realized that efficient and accurate maintenance wasn’t the only application for this technology. Something potentially more sinister was afoot. By June, the news had spread to consumer advice sites like Clark Howard and The Points Guy with warnings that this could be a way for Hertz to charge consumers with additional fees. In a May 20, 2025, informational piece, The Points Guy site cautioned:
“Of course, this also raises concerns that consumers could end up getting charged for more potential damages that a cursory glance from a rental car company employee may have overlooked in the past. We'll have to wait to see if that ends up being the case. So far, there have not been a lot of reports of issues like that.”
Finally, in the last few weeks, these concerns have been borne out. Numerous articles, capped by a piece in the New York Times a couple of days ago, have reported that Hertz, in particular (along with Dollar and Thrifty, which it owns), is using this technology to identify damage sustained during the rental and then to charge customers extra fees. Here’s a story about a particular customer experience from the New York Times piece:
“Kelly Rogers and her husband rented a car from Hertz at the Atlanta airport over the July 4 weekend to travel to a family wedding in Birmingham, Ala. The couple, who live in Scarsdale, N.Y., booked a minivan to shuttle family around, and the drive in both directions was uneventful, they said. When they returned the car in Atlanta, they inspected it and saw no damage. A Hertz employee inspected the vehicle upon its return as well, they said, and did not flag any damage.
But once the couple had passed through airport security, they received a notification via the Hertz app that its automated system had detected a dent in the passenger-side front door. They were charged $195: $80 for the damage and $115 in fees, including those incurred “as a result of processing” the damage claim and the “cost to detect and estimate the damage” that occurred during the rental. Hertz offered to reduce the charge to $130 if they paid within one day.”
The automotive site The Drive reported another narrative from a Hertz customer along similar lines, providing more details about the fees involved.
“A reader named Patrick recently rented a Volkswagen from Hertz’s location at Hartsfield-Jackson International Airport in Atlanta, which was in fact the first store nationwide to use the tech. When he returned the car, he did so with a 1-inch scuff on the driver’s side rear wheel. Patrick says he was alerted to the damage “minutes” after dropping the VW off, and with it, charges for the blemish: $250 for the repair, $125 for processing, and another $65 administrative fee. That’s $440 all told, for curb rash on one wheel.
This is all relayed to renters like Patrick over a web app, and while it’s easy to guess what the repair fee is for, the other two are a bit vaguer on the surface. Hertz defines the processing fee as “the cost to detect and estimate the damage that occurred during your rental.” The admin fee, meanwhile, “covers a portion of the costs [Hertz incurs] as a result of processing your claim.”
Thus, the valuation process (i.e., automated damage detection) is customer-specific, but the nickel-and-diming with the partitioned prices seems cookie-cutter, part of it related to the damage and part of it standard high-margin markup fees. This brings us to the distinction between established value-based pricing for segments versus the newer methods that, like this one, try to establish value-based prices for individual customers. The interesting question is whether they succeed and if they make sense for the business in the AI world.
Value-based pricing is becoming more fine-grained
To those of us processing the Hertz story from a pricing perspective, the evolution occurring in this story may have a familiar ring to it, even if the technological details are unique. We can understand Hertz’s logic as well as the reasons for customer and media outrage in response to Hertz’s new pricing approach almost instantly. A new technology is introduced to the marketplace and is embraced by a market leader. First, it is used for a different purpose (to save money by optimizing maintenance), then the company realizes that one of the technology’s applications, or even its most useful application, is to generate additional high-margin revenue through implementing more fine-grained price discrimination. The company then enthusiastically uses it to do exactly this, and in the process delivers customer-specific value-based prices. The price paid by every customer now corresponds to the value received by them on the technology-related dimension. The end result is that different customers pay different prices depending on the value received.
Whether it appears like it or not, Hertz is very much implementing a fine-grained or customer-specific form of value-based pricing by charging renters for damage to the returned vehicle during use. Different customers cause different levels of damage (or hopefully, none at all in most cases), so they pay different fees to cover that damage. The “economic value” received by the customer in the form of damage they cause, through reckless or careless driving, or sheer bad luck, is paid for with fees assessed by Hertz with pinpoint precision.
Two perils of technology-driven value-based pricing
I have written earlier about how value-based pricing is essentially a zero-sum game where the economic value (and even other types of value) generated by the seller’s offering must be split between the buyer and seller. If the customer receives the value, the seller loses it, and vice versa. In essence, from the seller’s standpoint, value-based pricing involves determining value to the customer and trying to take as big a slice of the pie as possible, before the customer greedily grabs it for themselves.
Many of today’s versions of value-based pricing don’t look like the value-based pricing that we may be used to. The “standard” value-based pricing approaches offer multiple tiers of products or services whose prices vary based on quality, targeted offers to specific segments (e.g., student discounts), or offers based on other constraints (e.g., happy hours, last-minute deals) to price discriminate based on valuation. Not to be confused with things like surge pricing (which still is a segment-based price discrimination), the more modern versions of value-based pricing price discriminate (or try to do so) at the individual level. The idea behind the Hertz pricing is that every single renter could potentially pay a different total price for a vehicle rental when the already existing variability in daily rental prices, insurance, plus numerous other optional extra benefits, are added together with the individual damage assessment2. It’s the old-fashioned good-better-best pricing as the base layer with a second layer of partitioned pricing of features and a third layer of technologically driven customized value identification to arrive at the final custom price for each renter.
For the Hertz case specifically, it is worth noting that the articles that have been written so far have mostly focused on customer outrage about the reliability of the technology being used and questioning whether the damage to the vehicle actually occurred. None that I’ve read have brought up the challenge of superior technology affordance or the challenge of perceived fairness, which I feel deserve greater attention because they are the key to understanding whether this pricing approach is sustainable or not.
The challenge of superior technology affordance. The AI-driven scanner that Hertz has adopted is able to detect damage to the vehicle that its human employees were not able to detect. This is an instance of charging a customer for a much smaller infraction that previously went unnoticed and unpunished because of superior technology affordance. In their dealings with companies as they consume products and services, customers perform minor and not-so-minor infractions all the time, virtually all of which go unnoticed and unpunished.
Someone may fill a plate at an all-you-can-eat buffet and not finish it, breaking its rule of paying for unfinished food. A parent may bring a 13-year-old to a concert and claim he is 11 to get the child ticket price. A fashionable Gen-Zer may buy an expensive outfit, wear it to a party, and then return it, taking advantage of the retailer’s lax return policies. A fast-food enthusiast may stock up on condiments, napkins, or utensils (or even toilet paper). These things cost money, but are ignored by the business and added to overhead costs, and marked up across the entire customer base. Basically, everyone pays a little more for the infractions of a few. This even applies to nefarious things like the loss from outright theft.
The Hertz move is significant because not only is the infraction defined in a more fine-grained way, but it is also penalized that way. On the face of it, this seems reasonable, if not for fairness.
The challenge of perceived fairness. What makes an effective pricing strategy so difficult to achieve is that perceptions of fairness of those who pay the price really matter. If customers perceive any pricing strategy as unfair, they are likely to respond with resentment, reduced loyalty, or even active backlash, such as giving negative reviews, social media campaigns, or even legal challenges. In the Hertz case, the shift to AI-driven damage detection disrupts long-standing consumer expectations about what constitutes reasonable wear and tear on a rental vehicle during use, and how that should be identified and remedied.
Like reference prices that set the baseline against which current prices are judged, customers’ past experiences with returning the vehicle set the baseline against which the new technology-driven method of damage detection is judged. Customers have historically operated under the assumption that minor dings, scratches, or scuffs that inevitably occur when a vehicle is driven will be overlooked, similar to airlines who don’t charge extra if the luggage is slightly heavier than the permissible limit or hotels that don’t charge extra for staying ten or fifteen minutes beyond the checkout time. This implicit social contract allows for a bit of flexibility, fostering goodwill and repeat business.
What’s more, charging individual customers now for what was a collective cost bundled into the price earlier makes the pricing method even more unfair. Many customers may feel like the rules have changed in the middle of the game without warning. The Hertz customers in both stories described earlier shared a keen sense that Hertz was charging them unfairly. Even further, despite claims of transparency, linking the minor damage to the actual fees charged (“Was the one-inch scuff on the driver’s side rear wheel really worth $440? Will Hertz really have to spend that much or are they gouging me?”).
Moreover, as I’ve already mentioned (I think this is an important issue for pricing decision makers in particular), there's an element of distributional fairness at play. Before Hertz adopted the new AI-scanner, the costs of minor damages were spread across the customer base by being added to the quoted rental rates, so that everyone paid a little more to cover the costs. Now, with AI-based detection, the burden shifts directly to individual customers, which sounds equitable in theory but can feel punitive in practice to those affected.
Imagine a scenario where a customer returns a car after a long trip on mostly gravel roads, only to be assessed a $500 fee for micro-chips in the paint that weren't even visible to the naked eye. This sort of thing is certain to erode trust, particularly among customers who are price-sensitive and budget-constrained (and even those who are not). They are likely to feel they are being unfairly taken advantage of by Hertz with sneaky pricing. What’s more, this sense of being mistreated is likely further exacerbated by the general belief among consumers that prevails at the moment that businesses are in a position of power and trying to exploit them3.
Mispredicting consumer response to the superior technology affordance-perceived fairness combination can sink value-based pricing
We have seen many versions of the Hertz story play out going back decades, with the same unhappy ending. Perhaps the most famous example is that from back in 1999, when Coca-Cola began testing a vending machine that could automatically raise prices when temperatures went up. (No machine learning necessary, simply install a thermometer in the vending machine and yoke the price to temperature4). While raising prices when it’s hot outside got all the buzz, the truth was that part of the pricing strategy being tested was also to lower prices when demand was expected to be lower.
At the time, Coca-Cola’s spokesperson told the New York Times that other modifications were also under discussion to adjust prices based on demand at a specific machine, and they were looking at “What could you do to boost sales at off-hours? You might be able to lower the price. It might be discounted at a vending machine in a building during the evening or when there's less traffic.” Of course, we know how that went. Six years later, David Leonhardt assessed the defunct vending machine project in this way:
It was not one of the great marketing moments in the company's history. In Internet chat rooms and newspaper editorials around the world, angry Coke drinkers denounced the idea. The word "gouging" got tossed around a lot. Pepsi gleefully accused its rival of exploiting consumers. Coke responded by running away from the heat-seeking vending machine as fast as possible. Company spokesmen said that Mr. Ivester [Coca-Cola’s CEO] was talking hypothetically and there were no plans to add a summer surcharge. Coke was actually looking for ways that vending machine technology could lower the cost of a drink, they added.
There are lots of other famous examples of technology-based pricing going wrong for the company. Back in 2000, there was Amazon’s failure with dynamic pricing, where “in the space of two weeks, online retailer Amazon.com has been forced to apologize, issue refunds and appease angry customers after it was found to have charged some people more than others in random price testing on its Web site.” Uber had to apologize and cap its surge pricing in Sydney after it raised prices during a hostage situation back in 2014. Last year, Wendy’s CEO had to clarify that they did not plan to introduce surge pricing (for burgers?) after mentioning that they were looking at dynamic pricing during an earnings call5.
All of these cases share a common narrative arc with the Hertz case, and they are also all likely to share the same eventual fate. Even though AI and numerous hardware innovations make fine-grained value-based pricing feasible and potentially lucrative, that doesn’t mean that it is a smart or effective thing to do. It will be interesting to see how long Hertz maintains this AI-based, customized, high-margin fee before it backs down and goes back to using its AI-scanner solely for its original intended purpose of optimizing fleet maintenance. The main lesson is that in the consumer space, moving from segmented value-based pricing to customized value-based pricing is a tough thing to do.
If you haven’t read the Pricing Conundrum recently, there’s a lot of new content you may have missed. I have written about Dishoom’s Matka promotion, what a 1963 case study about a planned price increase of a steam cleaner manufacturer teaches us today, the impact of financialization on valuation, and why retailers are failing at Black Friday pricing. There are also a number of new cases and teaching notes for instructors teaching pricing strategy. See Ironwood Resources, Streamline Energy Partners, the Houston Symphony Orchestra, and Maverick Drilling Technologies. A lot more is on the way.
Thanks for your support, and stay tuned!
Although the wording of the press release seems innocuous on first glance, you can see some ominous signs for their customers if you are reading closely. For example “transparent automated checks when picking up and dropping off their vehicles” should raise a red flag; when the word “transparent” is used by a company in the context of pricing changes in something like a press release, more often than not, it’s not good news for the customer.
The New York Times article reported a Hertz spokesperson as saying that so far (several weeks into its use at airport locations), less than three percent of rentals scanned by the AI scanner identified damage that Hertz charged its customers.
A lot can be written here about the widespread belief among consumers that companies are trying to use inflation as an excuse to raise prices and earn more profits over the past three or four years and its ramifications for pricing strategy. This is an interesting topic for another day and another post.
As the New York Times reported at the time, “The process appears to be done simply through a temperature sensor and a computer chip, not any breakthrough technology, though Coca-Cola refused to provide any details yesterday.
As many readers of this blog know, surge pricing and dynamic pricing are not synonyms. But that’s a discussion for another day.