AI Rewrites the Playbook for Reverse Logistics
As return rates remain elevated, margins compress and fraud tactics grow more sophisticated, retailers are being forced to rethink how reverse logistics fits into their broader business strategy.
Retailers have long known that the cost of a substandard returns strategy could tank both profits and customer retention rates. Consumers were expected to return nearly $850 billion in goods in 2025, or 15.8 percent of the year’s projected retail sales at the time, according to the 2025 Retail Returns Landscape report released by the National Retail Federation (NRF) and Happy Returns.
On top of that headwind, shoppers are getting picky and more demanding of a seamless returns process. Seventy-nine percent of online shoppers will abandon a purchase if the return policy does not meet their expectations, says DHL eCommerce Trends Report.
To protect margins and restore control, retailers are increasingly turning to artificial intelligence to guide decisions across the returns lifecycle.
At the NRF Rev 2026 conference on Jan. 11, Silicon Road Ventures’ founder and managing partner Sid Mookerji identified three core pillars where AI can optimize retail’s reverse logistics conundrums: fraud detection and prevention, predictive return forecasting and autonomous decision support on what to do with returned merchandise.
Related Stories
Although reverse logistics was once considered an afterthought at Target, according to senior director of returns and recommerce Lindy O’Brien, the mass merchant has turned over a new leaf from what she described as a historically “one size fits all” returns strategy.
“Everything was treated equally, and we knew it shouldn’t be,” O’Brien said.
The retail giant uses algorithmic decision-making to evaluate each returned item on its own merits, incorporating seasonality, historical demand, recovery rates and handling costs to determine whether inventory should be resold, donated, recycled or liquidated.
Store associates taking in returned items can follow simplified sorting instructions while the system determines the optimal downstream path, improving their decision-making capabilities.
“We knew that in the back room of our stores, we didn’t have room to do a bunch of sorts. We couldn’t ask them to use their labor in a taxing way,” O’Brien said. “Every single day, our algorithm is reupping what our recovery looks like, and helping give back-room information to store team members to sort inventory in a simple, binary way. That has made things seamless inside of our four walls.”
Not every retailer is deploying AI in headline-grabbing ways, but many are intentionally building the foundations that make the technology viable. Performance footwear brand Ariat International’s returns operation already functions as a rules-driven decision engine determined to correctly place each item that gets returned.
Sara Barber, senior director of supply chain inventory and returns at Ariat International, described a returns ecosystem that combines software and human decision-making to determine whether a product could be resold in a secondary market, an outlet store or a discounter.
“Behind the scenes, we have built a logic that says, if it’s apparel and marked ‘B-grade,’ it goes into this box versus that box. But if they marked it ‘A’, it does go back to the line, and the system also tells them which of our multiple DCs the product should go back to. So the business is in control of that logic. I can change that logic anytime. The associate on the floor just has to grade it, and then it tells it exactly how to move it to its final destination.”
Retailers walk a tightrope between easy returns and rising fraud
As retailers work to reduce friction for legitimate customers, fraud has emerged as the unavoidable counterweight.
Nine percent of returns are fraudulent, said NRF and Happy Returns’ survey of more than 350 e-commerce professionals. For Abercrombie & Fitch, fraud often exploits early refund programs and customer service escalation, according to Rob Sondergaard, the apparel retailer’s senior director of returns, fraud and abuse.
“Some customers that qualify for early refunds don’t end up doing the thing that you’re trusting them to do,” said Sondergaard. “You’ve got to keep an eye on who those customers are that qualify for early refund fraud, and which customers are trying to take advantage of a customer service agent. You have to make sure you have a team dedicated to looking at those times when you are losing money in the P&L and figuring out who those bad actors are.”
The NRF/Happy Returns report indicates that 85 percent of merchants say they are using AI or machine learning in their returns process to identify and combat fraud, but the results have been mixed. Only 45 percent of retailers believe these tools are effective at preventing fraud on their own.
According to Mookerji, “computer vision-based identification of merchandise is coming to returns” in a big way.
The UPS-owned Happy Returns is currently piloting computer vision-based technology of its own, deploying AI-powered returns auditing solution Return Vision to help retailers curb the issue. The returns management provider expects to roll the technology out to a broader audience in 2026.
According to Everlane, a participant in the Return Vision pilot, more than 85 percent of the company’s returns now happen in person through the Return Bar network.
The technology is aimed at flagging lookalike items that often pass off as the real product to the naked eye, in that they might have incorrect or missing logos, a strap that’s too wide, a mislabeled tag or fabric that doesn’t match.
As part of the pilot, flagged items are pulled for a rapid audit at one of the company’s regional Return Hubs. Using the Return Vision software, each product is photographed and compared against photos from the merchant’s online catalog. If the audit flags an inconsistency, the refund is withheld. If the item passes, the refund is released to the shopper.
The sustainable apparel seller says it has used the solution to confirm that more than 99 percent of items returns by shoppers are genuine. For the fewer than 1 percent of returns that are flagged, Happy Returns found that the average loss prevention amounted to $218 per fraudulent return.
FedEx, DHL expand returns infrastructure as retailers seek relief
Reverse logistics has become so pivotal that major logistics providers are dedicating specific branches of their business to shielding retailers from the returns onslaught.
Last summer, FedEx launched the FedEx Easy Returns box- and label-free service across 3,000 FedEx Office and Kohl’s locations across the U.S.
This program and others requiring no label or box are designed to further simplify the returns process for customers, who are catching on. Familiarity with no-label/no-box returns increased to 48 percent among returners in 2025, according to FedEx’s Returns Survey released in January, up from 37 percent in 2024. Usage increased to 41 percent from 31 percent in the year prior.
According to that survey, 66 percent of consumers describe the method as “convenient,” while another 67 call the process “stress-free.”
DHL is taking its own approach.
After bolstering its returns capabilities via its acquisition of Inmar Supply Chain Solutions to kick off 2025, DHL Supply Chain formally integrated the business with its own operations and technology to debut the DHL ReTurn Network in October.
The ReTurn Network is designed to allow DHL to serve a broader market across emerging brands and established retailers, without requiring customers to overhaul their existing distribution strategies.
The network expands DHL’s capabilities beyond dedicated distribution centers by offering 11 facilities to streamline returns processing for multiple customers and product types within shared centers.
DHL says its physical network helps clients save 10 percent to 50 percent of their returns costs.
“Our national network of shared facilities reduces warehousing, labor, and transportation costs while optimizing the handling and disposition of returned goods. It also provides customers with greater flexibility for managing fluctuating inventories and warehouse space during times of increased market volatility,” said Thomas Borders, vice president of operations for DHL Supply Chain North America’s ReTurn Network.
Like employees at Target and Ariat, workers in DHL’s reverse logistics network leverage a proprietary decision-making software called ReSKU, which is built to help standardize and hasten returns processing.
Content Original Link:
" target="_blank">
