The Case of the Inflated Appraisal
Updated: Oct 26
This article was originally published in INSTORE Magazine on July 3, 2026. Read the original here.
Sterling & Co. had built its reputation on precision and integrity, upheld daily by senior appraiser Elliot and store manager Marissa. Their expertise made the shop a trusted destination for collectors and sellers alike.
One afternoon, a new customer, Daniel, walked in carrying a limited-edition luxury watch in pristine condition. He explained he wanted to sell the timepiece, confident it would command a premium. He placed a printed report beside the watch: an online AI-powered appraisal tool had valued it at $28,400.
The number made Elliot pause, not because the watch was unusual, but because the estimate seemed off.
Elliot examined the watch and performed his own evaluation. He confirmed the watch was authentic. The serial number aligned with manufacturer records. Authenticity was not the issue. Value was based on current market demand, verified sales data, and condition, the watch’s true fair resale value was $14,800. Sterling & Co.’s buying offer, standard practice, was $9,500, leaving room for refurbishment, listing, fees, and risk.
The difference between the client’s expectations and reality was vast. The online estimate was inflated by more than $15,000, likely caused by the limitations of automated valuation tools.
The online appraisal generated an impressive-looking report featuring charts, predictive pricing curves, and automated resale forecasts. But while it looked sophisticated, the data behind it wasn’t rooted in the nuances of the current market. Automated tools often rely on broad datasets. Without visibility into how the algorithm calculated its estimate, it was clear only that the output didn’t align with Elliot’s professional evaluation.
When Elliot completed his evaluation, he presented Daniel with the store’s buy offer. Daniel instantly rejected it. Marissa tried to help by stepping in to explain how genuine resale values are determined, why market corrections matter, and how algorithmic tools often overestimate the value of luxury items when relying on incomplete data.
Daniel took his AI appraisal and his watch and walked out, accusing the team of trying to lowball him.
Within hours, a one-star review began circulating across digital platforms. Daniel accused the store of offering “less than half of what the watch is worth,” even claiming, “Sterling & Co. rips off their customers.” The review gained traction quickly, fueled by online debates about how valuations should be calculated.
Marissa and Elliot quickly called a meeting with the store’s owners, Henry and Lillian. In the meeting, Elliot demonstrated how easily the AI tool inflated valuations. Henry and Lillian agreed the store needed to get out ahead of the issue by responding, not with defensiveness, but with education.
Sterling & Co. launched a comprehensive educational initiative. They began planning a multi-part campaign covering topics like how true market appraisals are calculated, why algorithmic estimates can inflate values, and what factors influence resale pricing beyond what an online tool can detect. The goal was not to call out the customer but to empower the public with accurate information.
A key part of this effort was highlighting the credibility behind their process. Elliot, a GIA-certified appraiser, led the educational series to demonstrate what true industry-standard evaluation looks like. His GIA credentials gave collectors and serious buyers confidence in the legitimacy of what he was presenting.
Feedback was positive. Customers appreciated the transparency, and collectors began discussing the importance of expert examination over digital shortcuts.
Despite the success of their efforts, over the next month, two more customers arrived with watches they wanted to sell, all accompanied by AI appraisal printouts citing the same website. Each report estimated inflated values, despite the actual market ranges for those models. Each customer, like Daniel, expected Sterling & Co. to “match the online number.”
All of the watches were authentic. None were worth what the algorithm promised. A troubling pattern began to form.
The Big Questions
How should Sterling & Co. handle the growing number of customers arriving with inflated AI-generated appraisals?
What steps can Sterling & Co. take to protect their reputation when negative reviews are based on inaccurate appraisal practices?
Should Sterling & Co. publicly address the specific AI appraisal tool producing inflated values, or would calling it out risk escalating the situation?
SUMMARY OF RESPONSES: Educate before you offer: an appraisal isn’t a buying price, and AI estimates aren’t market reality. Most respondents say politely pass rather than insult sellers with a lowball, reset expectations upfront, and arm your website and social channels with content debunking inflated AI valuations.
Karl V.
Port Charlotte, FL
When a customer walks in with an AI-generated value, the store should kindly let the customer know that they are not interested in purchasing the watch at all. This would prevent the customer from being given an offer for considerably less than the AI-generated value. Not making an offer is better than an offer being perceived as a lowball offer.
Richard B.
Reading, MA
If you don’t know the client bringing the item in, do the following:
Educate the client before making an offer. First, identify the appraisal and who issued it. If it’s meant as an insurance appraisal that reflects current market value vs. MSRP on a watch like a Rolex Daytona, discuss that and the current market. If it’s an AI appraisal and inaccurate as to what a jeweler might pay, discuss that posture before making an offer.
If the item is one you really want to purchase and you suspect the customer’s expected value is not aligned with reality for an over-the-counter purchase, suggest he or she go to a couple of other stores, get their best offer first, then return to give you last look.
If they just want “out” of the piece immediately, make them an offer as you normally would.
Antonio S.
New London, CT
Appraisal values can differ from one place to the next. A jeweler could easily offer more if that jeweler feels it can move faster than another. Trusting AI has as much value as a listing on eBay or any other auction site. Always remind your client that what an item is appraised for has nothing to do with an offer to pay. If the jeweler feels the offer is so far away from what the seller believes they want, it’s best to walk away.
David A.
Rock Hall, MD
Always be truthful. Do everything you are able to do to prevent an accusation of bias or doubtful practices.
Use AI as a resource, not as an absolute fact finder. AI is getting smarter every day, but appraising correctly is not what AI can yet accomplish. AI still makes errors and occasionally offers biased results. Appraisers still need to conduct their own research to produce accurate, justifiable, and fair appraisal reports.
Make your reports compliant with USPAP and standards such as those required by NAJA and the other major appraisal organizations. Get your personal education in appraising not only from experience, but from conferences. classrooms and online courses. Become a “Certified” level appraiser to fully understand your duties to clients.
Steve W.
Poplar Bluff, MO
Do not make any offer on inflated value items. Explain to the customers that you in good conscience could not sell it for what they want for you to make any profit. Explain that the AI evaluation is really not based on actual market values. We would, though, put it out on consignment for what they want but warn potential buyers that the item is overpriced and give the buyer the opportunity to decide for themselves. We would tell the consignee that up front. If they want to leave it with us with that stipulation, that is their choice. If a customer wants to pay above the value, that’s their prerogative.
Meghan S.
Richmond, VA
One of the biggest breakdowns here is that Daniel’s expectations were never reset at the start. That needed to happen immediately. As soon as the AI report was presented, Elliot should have clarified that what Daniel had was not an appraisal but an automated estimate. That distinction matters. Without it, the entire conversation is built on the wrong foundation. That initial moment is also where Elliot’s credentials and process should have been introduced. Explaining how value is determined using verified sales data, condition, and current market demand establishes credibility before any number is discussed.
If Elliot had concerns about the estimate, those should have been communicated early. Setting expectations upfront avoids the perception later that the store is undervaluing the item. From a broader standpoint, this is an opportunity. A social media campaign focused on educating clients about the difference between automated estimates and professional appraisals reinforces that not all appraisers or appraisal methods are the same. One negative review, especially from a non-client, is unlikely to outweigh a strong, consistent reputation.
Shannon P.
Altus, OK
We would pass on it based on the appraisal brought in by the customer. We have had people do the same thing and bring in a store-inflated appraisal and just passed on it, knowing the customer expected a lot more than it was worth.
Jennifer F.
Colorado Springs, CO
AI is just going to be a part of expectations and bad information for a while. This is no different than the “gem scanning” apps that claim to be able to identify gemstones and teach you how to make lots of money buying jewelry at thrift stores and reselling to jewelers.
For my team: No matter what clients bring in for paperwork, we tell them we do our own analysis, and we do not want to see previous evaluations because we have to do market research. When we actually make our offer, we will pull up market data (gold pricing, resale values on eBay, new-condition comps) and tell them the offer we make has to be based on us accepting the risk of reselling and needing to make a profit if we successfully sell the item.
We have started encouraging clients with AI-generated information to take their items to the marketplace themselves, and if it doesn’t sell after three months to come back and see us for a realistic offer.
Klaus K.
Bristol, RI
My answer to the customer would be as follows: The main purpose of a jewelry or watch appraisal is for insurance; it gives you the full retail value needed to replace the item if it were ever lost or stolen. While it can give you a general idea of worth, it’s important to remember that an item’s true market value is simply what someone is willing to pay for it right now, and that number can be very different. An appraisal isn’t a price guarantee, and it’s up to you to find the right buyer. I know this might not be the answer you were hoping for, but I want to be upfront about how the market works. With all that in mind, based on your item, I can offer you X.
Megan C.
Port St. John, FL
Not to sound crass, but if a customer came into our store and acted that way about an unrealistic expectation from a random website, I would tell them they should go back to that site and expect them to purchase it for that price. Anyone can come up with what they think a fair value is, but that does not mean anyone is actually willing to pay that. I would obviously use more tact in saying that, but the reality is that’s like Schrodinger’s Appraisal — you can think it’s worth that until it’s time to sell and actually be faced with reality.
Tracy W.
Sierra Madre, CA
They were lucky that their customers listened to the long explanation. I find it better to keep it simple. When we are faced with customers having overly high expectations on price, we just say we don’t have the clientele for that at this time. Unfortunately, we can’t make an offer. It will be offensive to offer a realistic amount if they believe it to be more. We tell them we don’t want to insult them with a low offer. In the meantime, make sure your website has some short articles on the dangers of AI in appraisal valuations.
Ginger G.
Gladewater, TX
I usually tell customers that they can expect 10-25% of a written appraisal price when selling. That’s not a hard rule, but it gives people a good ballpark. It’s also a good idea to ask what they’re hoping to get if you are getting the impression that their price is too high. It is often better to politely pass than to make an offer much lower than they were expecting.
Marcus M.
Midland, TX
This is why I don’t buy from the public and this is why I can’t stand AI. I think Sterling & Co handled the situation correctly and tried to educate, but the American consumer has almost become impossible to educate because of AI. They’d rather trust what a computer is telling them than what an actual experienced human is. That’s sad. I think the only way to navigate the negative reviews is to keep trying to educate and maybe quit taking in appraisals like this. Also, probably a touch snarky but I would challenge someone who left a bad review to show me where they actually got the inflated price AI generated for them. I would guarantee that doesn’t exist.
Peter T.
Show Low, AZ
AI has NOTHING to do with this problem. Managing customers’ expectations is the issue. Customers often have inflated insurance appraisals for the items they want to sell. When a customer offers one of these appraisals, I review it and explain the difference in replacement appraisals (or falsely inflated appraisals) and the value they would receive if they sold the item. Most customers understand and either take the offer I give, or they reject it. But there are no hard feelings either way. There are always a few who refuse to accept reality. They can walk. I’ve never had one come back to pester me either in person or in reviews.
Mark S.
Plantation, FL
We look at what the client is offering for sale. If it is something that we don’t need, we don’t even make an offer. We don’t want to insult the seller by making a lowball offer and have them leave the store the same way their client left. If this item can be of use to us, we make an offer that is fair and reasonable to both parties. Most of the time, our offer is accepted. We don’t want to insult the seller, and we politely tell them that we are not interested in what they have to sell.

Comments