AI in resale and second-hand fashion: authentication, pricing and listing

KEY TAKEAWAYS Summary by the editors

1. Resale is a segment of single items: each product must be identified, authenticated, graded, priced and listed individually, so the cost of handling one item decides profitability.
2. McKinsey and The Business of Fashion forecast that the second-hand fashion market will grow two to three times faster than the firsthand market through 2027.
3. The RealReal reported that 85 percent of items were launched with AI pricing assistance by the end of 2024 and planned for its Athena image recognition tool to process about half of items by the end of 2025, according to PYMNTS.
4. eBay's AI listing tool generates titles, descriptions and category data from product photos; eBay reported in 2023 that over 95 percent of testers adopted the AI-generated descriptions.
5. AI can support authentication, but high-value items still need expert checks, clear accountability and a process for disputes.

AI in resale and second-hand fashion is used to identify items from photos, support authentication, suggest prices, write listings and improve search for one-off products. It matters more here than in most segments because every item is unique and margins depend on how cheaply and accurately each piece can be processed, so automation directly decides whether resale is profitable.

Why does AI matter differently in resale fashion?

A firsthand retailer creates product data once and sells many identical units. A resale business does the reverse: it handles many units that are each a product of one. Every item needs identification (brand, model, size, material), a condition grade, a price that reflects rarity and demand, photography and a description. For luxury goods, authentication adds another costly step.

The segment is growing. McKinsey and The Business of Fashion forecast in The State of Fashion 2026 that the second-hand market will grow two to three times faster than the firsthand market through 2027. ThredUp's 2026 Resale Report, based on research by GlobalData, sizes the global second-hand market at about 393 billion US dollars, roughly 10 percent of total apparel spend, and projects the US market to reach 78.8 billion dollars by 2030. ThredUp's report also describes AI as the engine scaling resale by improving search, discovery, pricing and authenticity checks.

What are the main AI use cases in resale?

Read also

AI in sportswear: how it shapes performance product, community and demand

How does AI create listings for second-hand items?

Listing is the most labour-intensive step for individual sellers and for platforms that process consigned goods. eBay described its magical listing tool in September 2023: a seller uploads a photo in the app, and the system generates a title, description, product release date and category, and can add pricing and shipping suggestions. eBay reported that roughly 30 percent of US app-based sellers were trying the feature daily by late July 2023 and that over 95 percent of testers adopted the AI-generated descriptions.

The value is less about writing and more about structure. A generated listing with correct brand, size, material and category is easier to find, compare and price. The weakness is accuracy: models can misread a label, confuse similar models or overstate condition, so sellers and platforms need a review step.

How does AI price and authenticate resale items?

Pricing a unique item means estimating demand from comparable sales, condition, rarity, season and current trends. According to PYMNTS reporting on The RealReal's results in February 2025, the company uses an AI-driven pricing engine and said 85 percent of items were launched with AI pricing assistance by the end of 2024. It also described Athena, which uses image recognition to support authentication and pre-populate listing attributes, and expected it to process about half of items by the end of 2025. The company reported that automation reduced processing time by over one full day.

Authentication is where AI is most useful and most sensitive. Image models can flag inconsistencies in stitching, logos, hardware or materials and route doubtful items to experts. They should not be the sole decision-maker for high-value items, because a false approval damages buyer trust and a false rejection harms honest sellers.

What data does AI in resale need?

Resale AI depends on data that firsthand retail rarely has to manage:

• Reference catalogues: images and attributes of original products, ideally from brands, to identify models and seasons.
• Transaction history: realised prices, time to sell and condition, as the basis for pricing models.
• Authentication records: confirmed genuine and counterfeit examples, with expert notes, to train and test models.
• Condition standards: a consistent grading scale with example images.
• Return and dispute data: to spot listings where AI descriptions or grades were wrong.

Brands running their own resale or take-back programmes have an advantage here: they hold original product data and imagery, which makes identification and description far easier.

What risks are specific to resale?

1. Counterfeits passing checks: sophisticated fakes evolve, and models trained on older examples may miss them.
2. Misdescription: AI-generated listings that overstate condition or misidentify a product lead to returns and disputes.
3. Pricing feedback loops: if many sellers follow the same algorithmic suggestions, prices can drift away from real demand.
4. Seller trust: consignors may distrust automated valuations that they cannot understand.
5. Brand concerns: brands may object to how their products are described or priced on third-party resale platforms.

Read also

Resale and recommerce: how second-hand reached the brands

What should a resale business or brand do first?

Identify the step with the highest cost per item, usually listing or authentication, and automate the routine part of it while keeping expert review for exceptions. Build a clean dataset of realised prices and confirmed authentication outcomes, because that data, not the model, is the long-term advantage.

Brands entering resale should start with their own product master data: images, attributes and original prices by style. That makes AI identification and listing far more reliable and keeps the brand's voice in second-hand channels. Measure success by processing cost per item, time to sell, sell-through and dispute rates.

Frequently asked questions

How is AI used in second-hand fashion?

AI identifies items from photos, generates listings, suggests prices, supports authentication and improves search for one-off products. eBay uses AI to generate listings from photos, and The RealReal uses AI for pricing and image recognition. These tools reduce the cost of handling each individual item.

Can AI authenticate luxury goods?

AI can flag inconsistencies in logos, stitching, hardware and materials and route suspicious items to experts. For high-value goods it works best as triage alongside specialist review, because counterfeits evolve and wrong decisions damage trust.

How do resale platforms price used clothing with AI?

Pricing models compare realised prices of similar items, adjusting for condition, rarity, season and demand. The RealReal reported that 85 percent of items were launched with AI pricing assistance by the end of 2024. Quality depends on large volumes of reliable transaction data.

How big is the second-hand fashion market?

ThredUp's 2026 Resale Report, based on GlobalData research, sizes the global second-hand market at about 393 billion US dollars, roughly 10 percent of total apparel spend. McKinsey forecasts second-hand fashion to grow two to three times faster than the firsthand market through 2027.

GuideThe complete guide to AI in fashion e-commerce, marketing and retailRead the complete guide