Resale Platforms Wire Inventory Into AI Shopping Agents

Vestiaire Collective is opening its pre-owned luxury inventory to OpenAI's agentic shopping systems, a move that could place secondhand Prada alongside new-season stock in AI-generated recommendations. The plan, outlined by CTO Rémi Bouchez, appointed in March, follows the platform's quiet launch last month of a dedicated app plugin within ChatGPT that lets shoppers search and compare pre-owned listings through conversational prompts — including side-by-side comparison of up to three listings on price, condition and seller information.

For resale platforms, the commercial stakes are measurable. The RealReal introduced its own AI shopping assistant, Ask TRR, built on Google Gemini, last month; CEO Rati Levesque told the company's third-quarter earnings call that Ask TRR's early adopters spend 20% more per transaction than non-users. The RealReal is also restructuring data across its 1.5 million listed items to make them easier for external AI agents to find and recommend — a preparation for shopping journeys that begin off-platform.

Depop has enabled search through Apple's Visual Intelligence on supported iPhones. Bouchez argues mainstream adoption of agents such as Muse, Instinct and Duet would reshape how luxury shoppers weigh options. "These agents are capable of ingesting a ton of information — way more than human shoppers can," he said, noting agents can flag discreet terms, questionable seller behaviour and authentication details before recommending the best-value option.

How much does listing quality move the needle?

Vestiaire's partnership with AI photography company Photoroom, announced last month, standardises the main photograph of every new listing — roughly 40,000 items a day — by removing backgrounds, adjusting lighting, centring products and adding shadows. A randomized A/B test conducted by the two companies found processed listings achieved a 5% higher sell-through rate than those using Vestiaire's previous in-house editing tool, with unique product page views up 3% to 7%.

Vestiaire has since measured an approximately 6% improvement in the proportion of enhanced listings selling within seven days, without a corresponding rise in returns. Bouchez says the platform is now experimenting with more ambitious enhancements such as "digital ironing" — technology that risks undermining the trust resale depends on if it erases defects like scuffs.

"Trust is a big pillar for Vestiaire — we invest a lot in that trust, on manual verification, on assessing a listing before it's published," Bouchez said. "We want to keep the same level of trust but lower the investment with AI."

What determines ranking on resale marketplaces?

Vestiaire uses a three-stage ranking system: identify relevant products, reorder by marketplace factors, then adjust to individual preferences. Those factors include condition, pricing, location and seller reliability. Proximity matters on a global platform — nearby items mean faster shipping, fewer customs complications and easier returns. Vestiaire assesses fair price against condition, location and whether the seller is a professional offering returns; an implausibly cheap luxury item raises authenticity questions rather than bargains. Bouchez says the analysis remains a human-AI combination.

Depop's algorithm prioritises query match across brand, category, size and colour, alongside popularity signals such as views, likes and bag additions, plus seller history and location. For underperforming listings, Depop's guidance is new photographs, fuller descriptions and a pricing review.

Poshmark's "Smart List AI" generates titles, descriptions and categories from uploaded photos, and automates offers to buyers. Chief product officer Heather Friedland said listing creation is now five times faster and selling 50% faster. "We want sellers to focus most on what they do best: source great items and have Poshmark take care of the rest," she said.

Can sellers and shoppers still influence discovery?

Depop chief product officer Sonia Biddle says shoppers increasingly start in the personalised Suggested For You feed rather than search, with AI identifying listings by category, colour and style even when keywords are imprecise. "Fashion often starts with an idea: an era, an occasion or someone whose style you love, and that's not always something you can put into words," she said. Keyword hashtags may matter less; photography, pricing and reliability more. Depop's visual search accepts photos or screenshots — including of first-hand retail products — to surface similar secondhand pieces.

Shoppers, Biddle noted, are deliberately "training their algorithm" through likes and saves. To counter personalisation's echo effect, Poshmark launched a "Tastemaker" programme in September with fashion figures including Courtney Grow, Colleen Allen and Jalil Johnson curating shoppable edits, and staffed a dedicated human merchandising team alongside its AI specialists. "On trends, a purely algorithmic system tends to amplify what's already popular," Friedland said. "Human curation is our counterweight."

What does this mean for first-hand luxury brands?

Resale platforms are accumulating long-horizon demand data that brands' own sales cannot capture. "Resale tells brands something their own sales data can't: what people still love years after it left the shop floor," Biddle said. Poshmark tracks patterns across searches, likes and transactions to separate fleeting trends from enduring desirability.

If AI assistants surface archival products as readily as current collections, luxury brands may compete with their own back catalogues without controlling presentation or comparison. As agents increasingly decide which listings reach a shortlist, the open question is how much influence brands, sellers and shoppers retain over what the algorithm ultimately recommends.