← All retail signals
Trust signal

Proving You Actually Wore It

A returned dress can now be checked against a fraud model before the refund clears, and retailers are betting that scrutiny pays for itself.

What is it?

Retailers are pairing AI fraud-detection with near-universal return fees, turning returns from an absorbed cost into a screened, priced transaction. Shoppers now use the same kind of tool to draft the claims retailers are trying to catch. Whether this is fraud control or margin recovery in disguise stays open.

Why now?

Trust between shopper and retailer frayed as free, unquestioned returns became a channel for abuse cheap enough to run at scale. AI cut the cost of both sides — generating a claim, and flagging one. The same capability now sits on offence and defence. Neither side holds a lasting edge.

Who’s building it?

Fraud-detection vendors and retail loss-prevention teams, working alongside the pricing desks levying the new fees.

Fraud-detection vendorsLoss-prevention teamsRetail pricing desksWardrobing shoppersPayment processors
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — fraud-detection vendors selling screening tools into retail loss-prevention teams.
Buyers — major retail chains adopting both AI screening and paid-return policy together at scale.
Gatekeepers — payment processors and regulators setting what fees stay allowed.
What’s at stake
Retailers recover margin lost to wardrobing and fraud, but risk alienating the honest shoppers who return in good faith.
Detection accuracy decides who bears a false accusation.
Fee-and-screen policy becomes a brand differentiator, not background cost.
What’s unsettled
Whether detection can keep outpacing the AI-generated fraudulent claims it targets.
Whether fees deter fraud or simply tax honest returners.
Whether returns policy stays a competitive lever, or converges into an industry norm.
02 · How the demand matureda progression, not a timeline
Screening
Detection systems flag suspect wardrobing patterns before a refund is ever issued.
Pricing
Fees spread across most major chains, pricing in the cost of abuse.
Adaptation
Shoppers and fraud tools each adjust to the other’s latest move, in a loop that keeps resetting.
Norm
Screened, priced returns settle into the default shopper experience, no longer treated as unusual.
03 · Where the market goesthe forks, not a prediction
→Consolidation or fragmentation — large fraud-detection vendors become standard infrastructure, or every retailer builds its own bespoke screening layer instead.
→Fee structures could converge into an industry norm, or stay wildly retailer-specific for years.
→Shoppers may keep adapting their tactics faster than detection models can retrain against each new one.
→Regulation could cap fees before the market settles on its own terms.
Everything here turns on generalisation: whether AI can tell a genuine return from a manufactured one reliably enough to justify the whole system.
04 · The Verification Stack deskadjacent markets, moving now
The Verification Ritual
Platforms add human checks as bot traffic grows.
Reddit adds human checks for accounts that show signs of automated behaviour.
Biometric Update
Proof at Capture
A new testing framework judges deepfake detection claims.
A world-first testing framework for deepfake detection.
GOV.UK
Human, Verified
AI-enabled fraud has entered official crime counts.
The FBI’s 2025 report counted 22,364 complaints involving AI, with $893 million in losses.
The Register
05 · The Wire
01
AI risk scores and image checks screen returns before refunds
CBS News ↗
02
Shoppers use AI too.
Modern Retail ↗
03
72% of US retailers now charge for some returns, NRF finds
Marketplace ↗
The panel on this signal
They agreed on
–Fraud detection and fee policy have merged into one returns strategy, no longer separate cost lines.
–The technology arms race has no clear leader.
–Detection accuracy, not fee size, decides where shopper trust ultimately breaks down.
The retail analyst disagreed on
–Fraud-screening fees on returns generate revenue regardless of whether flagged transactions are genuinely fraudulent; that overlap makes it difficult to separate a fraud-prevention rationale from straightforward margin recovery applied across the return category.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.