Point of Sale + AI for Thrift Stores and Resale

TLDR: A resale point of sale fails when staff guess the brand, the size, or the price. Walk every item through identify, price, tag, and checkout on one record, and use AI only where that routine is slow. Staff still check the brand, the size, and the price before that tag hits the rack.

What a resale register has to do

A grocery register looks up a barcode that already exists. A thrift or resale register usually meets the item for the first time. Someone has to name it, price it, print a tag, and sell it later without creating a second record. If those steps live in different notebooks, the floor count and the drawer will not match.

Point of Sales + AI is that path: intake, an in-store price, a production tag, and checkout on the same inventory. You can run the routine by hand. AI is there to do the lookup and the price recommendation so a new hire does not invent them.

Run intake the same way every time

Give the person at intake a fixed order. Skipping a step is what creates the mystery tag that nobody can ring up.

1. Look for a brand mark, a size tag, and a material or care label. If the label is gone, say so. Do not guess a brand from the color.
2. Name the item in plain words: denim jacket, fleece pullover, cast iron skillet. Add who it is for and the size.
3. Note color, material, and condition in one line. Condition is what a shopper can see: stain, hole, missing button, light wear.
4. Put it in a department the floor already uses, such as women's outerwear or home and kitchen. New department names make reporting useless.
5. Set a price before the item leaves the table. An unpriced item does not belong on a rack.

Photograph anything you might list later, even if it stays on the floor today. One clear front photo, the brand tag, and one flaw is enough. You will not get a second chance once it is on a hanger.

Price it for this store, not for the internet

Ask what this store sold last time, not what a listing is asking online. Asking prices are wishes. Sold prices are evidence. If you do not have your own history yet, use recent sold comps for the same brand, item, and size, then mark the price down until it fits your floor. A $40 jacket that sits for a month was priced for a different store.

Write the reason next to the price in the first weeks: "sold comps around $32, floored at $20." After a month you can see which reasons were right. That note is also what you hand an AI later, so it learns your store instead of a national average.

Decide the floor versus online before you tag

Most goods should stay in the store. Shipping and marketplace fees eat the recovery on a $15 shirt. A simple test is enough: if the online sold price, minus fees and the cost of packing, is clearly higher than the floor price, pull it for the website and eBay. If it is close, keep it on the floor. Speed of sale matters as much as the sticker.

Tag the floor item with brand, item, size, color, department, and the in-store price. The tag is the contract with the register. If the tag and the record disagree, fix the record before the item hits the rack.

Where AI takes the same steps

HAMMOQ Point of Sales + AI runs this routine from a photo or a barcode. It fills brand, item, size, color, and department, then recommends an in-store price for that location. Staff still accept the tag. They are not asked to start from a blank screen, and they are not locked out of a correction when the photo is a generic black shirt.

Checkout stays on that record. When the item sells, the floor count drops with the sale. High-recovery items can move to inventory for eBay and your own website. Poshmark and Mercari are not connected yet.

Mistakes that keep showing up

• Pricing from the tag the donor left on, or from a new retail price.
• One department called "misc" that becomes half the store.
• Pretty photos and no size or brand tag in the frame.
• A price in the register that was never printed on the item.
• Sending low-dollar goods online because the listing looked more professional than the rack.

Walk two items through it

Treat these as practice, not as prices to copy. Your own sold history replaces the numbers.

Denim jacket. The collar tag says Levi's, women's, medium, cotton. The name on the tag is a medium-wash jean jacket, not a style you guessed from the wash. Recent sold comps sit around $32. This floor tags it at $19.99, because a jacket in that range often sells in the building before a shipment is worth the fee. The printed tag and the register record both show brand, item, size, color, department, and $19.99.

Cast iron skillet. The mark on the bottom says Lodge, 10.25 inch. There is no clothing size, and it does not go in a department called misc. It goes to housewares. Sold comps around $31 are the online evidence. A floor tag of $22.99 fits when someone in the store will buy it this week. If the bottom is cracked, write that, or do not sell it. A clean mark does not cancel a crack.

Run the next ten research items this way before you change software. If those ten tags are complete and each price has a reason, you are ready for a draft. If they are not, a tool will only fill the blanks faster.

If you want to see these steps run with AI on your own items, book a demo and we will walk through them with you.