I stood in front of a Value Village tag and felt a small shock. A vase that once sold for CAD 3 (€2) new carried a CAD 8 (€5) sticker in the shop window. That moment makes you wonder who’s deciding value now.
I’m going to walk you through what Savers, the chain behind Value Village and Unique, says it’s doing with an AI called ThriftIQ, why managers are breathing easier, and why shoppers are not. You should know the AI replaces the work of employees who used to determine prices, and that matters for both workers and buyers.
At one store a ceramic vase carried a surprising tag — how did thrift pricing work before AI?
For years, pricing at Savers relied on human judgment: team members graded condition and quality and applied category price rules. Jubran Tanious, Savers’ president and CEO, explained that two employees could look at the same shirt and give different prices because condition calls are subjective.
That variability drove the push for automation. Rather than asking workers to grade condition, ThriftIQ asks them to identify the brand and lets algorithms combine brand, seasonality, and sell-through data to pick a price. You hear the efficiency pitch, but you also hear a shift in power — from human hands to a dataset.
Will AI make thrift stores more expensive?
Short answer: it can, but the company insists it won’t. Savers has faced criticism — notably from CBC in Canada — after shoppers found inflated tags (the vase example above). The worry is simple: if an algorithm is tuned to chase profit, prices can creep upward. On the other hand, Savers executives claim pilot stores showed average prices that were the same or lower than the rest of the fleet and still kept items at roughly 40% to 70% off traditional retail.
On an earnings call, executives pointed to better profit and faster ramp-up — what did the numbers say?
At the recent earnings call, CFO Michael Maher and CEO Mark Walsh said ThriftIQ helped pilots deliver higher sales yield and gross profit. Savers projected net sales of $1.77 billion (€1.63 billion) to $1.79 billion (€1.65 billion) for the back half of the year, and reported comparable-store sales growth of about 3%–4%.
Walsh argued the benefit wasn’t higher tags across the board but smarter pricing and simpler operations. To me, ThriftIQ is a thermostat for prices: it’s meant to stabilize, not spike, but the settings matter. The company says ThriftIQ raised gross profit dollar growth by about 100 basis points in pilot stores versus non-pilot stores.
How does ThriftIQ set prices?
According to Savers and reporting from CNBC, ThriftIQ was built with Kaizen Analytix using Savers’ proprietary data. The inputs reportedly include brand identification, seasonality, and sell-through rates. The team member’s role narrows: name the brand, scan the barcode, and the system suggests the price. That removes subjective condition grading from the frontline process.
In Canada shoppers complained loudly — are consumers worried about surveillance pricing?
Customer pushback matters. Surveys from May show about 68% of Americans worry surveillance pricing will raise costs; when a large chain automates price decisions, that anxiety intensifies. Savers is mostly owned by Ares Management, and the company has had to answer whether the algorithm will change tags after they’re printed.
Walsh told reporters the tags don’t change once printed and that the company is not using dynamic or surveillance pricing. Still, shoppers and privacy advocates point to the broader trend where AI and data can target prices based on behavior — turning a shopping tag into a magnifying glass on our buying choices.
Is Savers using dynamic pricing or surveillance pricing?
Savers’ public stance: no. Walsh said ThriftIQ isn’t being used for dynamic pricing and that “once those garments are priced and tagged, that tag doesn’t change.” The company emphasizes pilot results showing better sell-through and larger baskets without raising average prices versus the fleet. But reporters from CNBC and the CBC have pressed on what proprietary datasets and models power the tool, and shoppers remain skeptical.
I’ll say this plainly: you should watch how pilots scale. If ThriftIQ spreads to dozens more stores, you’ll start seeing patterns — whether prices cluster upward on certain brands, or whether items actually move faster and for less. Who watches the price-watchers when a private-equity-backed chain leans on machine learning to score a sweater — you or them?