# Dave Renfrew, Independent Estate Sale Picker at Renfrew Resale (self-employed, 1 person + weekend help from my brother-in-law) — read of Antique Valuator AI, June 5 2026

> 9 years picking estates and antique malls in the Midwest. I clear roughly 3-4 sales per weekend in season and resell through a mix of eBay, two local antique mall booths, and Facebook Marketplace. I own a 2019 Transit that I have never once called anything but "the van."

## How I got here

Googled "antique appraisal app iphone" at around 10pm on a Tuesday after a bad week. I bought what I was pretty sure was a legit McCoy bowl at a Rockford estate sale, got home, looked closer, and it's a knock-off. Paid $28, worth maybe $6. I've been using WorthPoint for three years and Google Lens before that and I still get burned maybe once a month. I was looking to see if anything new had come out. This page was not the top result. I clicked it maybe four or five results down because the title had "estate sales" in it.

## What I clicked first

The hero line "Find Hidden Treasure in Estate Sales" is the right sentence. That's genuinely what I am doing on a Saturday morning at 7am with a flashlight in someone's basement. I didn't roll my eyes at that. I scrolled.

Then I hit "Instant Photo Appraisal" and "Our AI analyzes condition, style markers, demand signals, and comparable sales to estimate real market value in seconds." That's a large claim. I read it twice. I didn't disbelieve it exactly, but I've seen a lot of apps make that promise and then give me a number that's off by 300%.

## Where I paused

The "Detailed Condition Report" section. Specifically: "wear patterns, restoration history, authenticity indicators, and what matters most to collectors of that specific piece." That last clause is the interesting one. Because what matters to a flow blue china collector is completely different from what matters to someone hunting mid-century Heywood-Wakefield furniture. If the AI actually knows category-specific tell signs, that's worth something. If it's giving me the same boilerplate for a Tiffany lamp and a Depression glass butter dish, it's not useful. The page doesn't tell me which.

## What I distrusted

The bottom of the page killed some of my momentum. "Built by Wishdeal Studio. Browse Free. Unlock for $5. Adopt for $99." I had to re-read that a few times. This is not a product I can download. This is a studio selling the idea of a product to someone who might build it. "Estimates only, no live customer revenue claimed" is at least honest, but it also means I cannot try this thing. There is no app. There is no demo showing a real appraisal. The "Try it Live result" section on the page appears to be a mockup, not an actual output.

So I'm reading features for something that does not exist yet. Every feature description is written in present tense ("See recent sales," "Walk into estate sales armed with knowledge") which reads like a live product until you get to the fine print. That's a trust issue for me, not because they're hiding it, but because the framing made me feel like I was evaluating a tool, not an investment pitch deck.

Also: no mention of what categories the AI covers, no accuracy data, no information on training data sources. WorthPoint has 600 million price results in their database. I don't know what this would have.

## What would convince me

An actual output. Not a screenshot of a nice UI. A real photograph of something genuinely ambiguous, like a piece of unsigned pottery or a piece of furniture with conflicting signals, run through the tool, with the result shown including where it was wrong and why. That kind of transparency would make me take this seriously.

Or: someone like me writing a paragraph about how they used it at an actual sale. Not a testimonial headshot. A specific story. "I photographed a console table in bad light in a garage and it correctly flagged it as a 1940s Widdicomb piece and said comps were running $800-1100. I paid $75." That sentence would be worth more than every bullet point on this page.

Accuracy percentage on a real test set would also help. Even a rough one. "We ran 200 eBay sold listings through the model blind and hit within 20% on 140 of them." Something.

## What I'd ask in an email reply

1. What categories does the model perform best on right now, and what does it struggle with? (I deal in a lot of American art pottery and vintage furniture. If those are weak spots, I want to know upfront.)

2. Does it handle regional pricing variation? A piece of signed WPA art sells for different money in Chicago versus rural Nebraska. Is the comparable data pulling from national averages or can it be filtered?

3. Is there actually a working prototype I can test, or is the $5 buying me a build plan?

## Verdict: on-the-fence

The features described are real problems I actually have, and "Smart Negotiation Tips" and "Selling Guidance" are the two that would genuinely save me money or time. But I can't evaluate any of this because there's nothing to use yet. I'm not dismissing it, I'm just not sure who I'd be emailing or what I'd be buying into.

---
*Memo by skeptic persona, generated 2026-06-05. Studio breaks own self-grading loop.*