Cross-border recommerce is becoming a larger part of digital commerce in Asia, but the model depends on more than matching buyers and sellers across markets. When secondhand luxury goods, collectibles, fashion, and culture-linked merchandise move across borders, buyers must trust items they cannot inspect in person, sellers they may never meet, and platforms operating across different jurisdictions.

This is especially important as resale activity moves into higher-value categories. Luxury bags, watches, fashion pieces, K-pop merchandise, and other collectible items increasingly carry asset-like value, which also raises the incentive for counterfeiting. Visual inspection alone can struggle against better-made fakes, while buyers also need confidence around shipping, payments, item condition, dispute handling, and the total cost of cross-border purchases.

For recommerce platforms, authentication is becoming part of the infrastructure that allows transactions to scale. AI-assisted analysis, material-level testing, item-level data, human expert review, and authentication records can help reduce uncertainty across the transaction chain. At the same time, better logistics, familiar payment options, and partnerships with industry or public-sector stakeholders can help turn verified inventory into a more reliable cross-border experience.

In this TNGlobal Q&A, Jaewha Choi, CEO of Bunjang, discusses how authentication, buyer experience, and platform trust are shaping the next phase of cross-border recommerce. Choi explains how Bunjang is applying Corelytics, its authentication system, to support luxury verification, why human expertise remains central to AI-assisted authentication, and how resale data is beginning to influence culture-driven categories such as K-pop.

Cross-border recommerce has grown quickly, but trust remains a major barrier for buyers. Why is authentication becoming more important as secondhand luxury and collectibles move beyond domestic marketplaces?

Cross-border recommerce scales not because people like secondhand, but because people trust it.

Luxury recommerce has been expanding nearly 15 times faster than the overall luxury market, and secondhand items are increasingly treated as assets rather than keepsakes. One example was a signed single from a first-generation K-pop artist that recently traded for more than US$2,700 on Bunjang Global, nearly two decades after its release. As prices rise, so does the incentive to counterfeit. Today’s counterfeits are sophisticated enough that even luxury brands sometimes fail to detect them; there have been cases of brands unknowingly repairing fake items.

Distance further compounds the problem. Within a domestic market, buyers have informal trust mechanisms. For example, they can inspect an item in person, meet the seller, and rely on local reputation and recourse. However, the moment a transaction goes global, these are no longer feasible. The buyer is committing money to an item they have never touched, from a seller they will never meet, in another jurisdiction. Trust must come from somewhere else, and that somewhere is authentication.

In our view, without trusted authentication, there is no such thing as cross-border luxury recommerce.

Bunjang has described Corelytics as part of its authentication process. What has the company learned from scaling inspections from tens of thousands of items, and how has that changed the way authentication is handled operationally?

Every inspection makes the next inspection smarter. Our inspection volume grew from more than 5 million items in 2024 to more than 8 million in 2025, and we expect to surpass 10 million items in 2026. Because Corelytics is built on continuous data input from every new inspection, each item analyzed makes the system sharper. Today, our authentication accuracy stands at close to 99.99 percent.

Many authentication systems still combine expert human review with AI-assisted analysis. Where does AI add the most value today, and where is human judgment still essential?

AI is most valuable where human expertise reaches its natural limits, not where it replaces people. In the case of Corelytics, it adopts non-destructive X-ray fluorescence and digital microscopy to capture what we call an item’s material DNA, or trace-element compositions that reveal differences invisible to the naked eye. Our model, trained on hundreds of thousands of authentic and counterfeit data points, renders a determination in just a few seconds. Today’s superfakes are crafted with high-quality leathers and carefully matched hardware precisely to defeat visual inspection. Material-level analysis is where they can no longer hide.

That said, human judgment remains essential at both ends of the process. Our authenticators conduct up to 18 inspection stages across roughly 154 checkpoints before scientific analysis begins, and the architecture is deliberately human-in-the-loop. This essentially means our experts handle edge cases, contextual signals, and final accountability.

The future of authentication is not AI versus experts. It is AI because of experts.

As inspection data grows, how can recommerce platforms improve accuracy while managing edge cases, rare items, changing counterfeiting methods, and false positives?

If your system is not learning, scale is meaningless. More inspections without better intelligence simply create more work. Where authentication relies on visual comparison and an inspector’s experience, higher volume adds workload faster than it adds capability, and rare items compound the problem because the pieces most worth faking are the ones inspectors have seen least often.

For Bunjang, three principles guide us as we improve and fine-tune the accuracy of Corelytics’ authentication process.

Physical evidence beats visual judgment. Counterfeiters evolve quickly, but authentication has to stay one step ahead. Anchoring authentication in material science means our accuracy does not erode each time counterfeiting methods evolve cosmetically.

Every inspection must improve the next one. Corelytics is designed so every inspection, including rare items and edge cases, feeds back into the model. Because the system operates independently of brand-provided data, coverage is not limited to what manufacturers choose to share. It grows with what actually passes through the market.

Errors are data, not failures. We treat every misclassification as a training signal for the system. Moreover, the human-in-the-loop design means Corelytics escalates rather than forces a determination, enabling us to manage false positives without sacrificing throughput.

You have referred to authentication as a kind of trust infrastructure. What does that mean in practice for buyers, sellers, logistics partners, payment providers, and platforms?

Trust should not have to be rebuilt at every step of a transaction.

For buyers, it means confidence before purchase, not only after delivery. Every cross-border order on Bunjang Global undergoes a final quality inspection in Korea, with buyers notified of the item’s condition before it ships. For the highest-value items, our BGZT Luxury store pre-authenticates every product through Corelytics before it is listed. For sellers, authenticated items translate directly into liquidity: items sell faster, command fairer prices, and generate fewer disputes.

For third-party logistics and payment partners, disputes, returns, and chargebacks have always been areas of friction. When the item at the center of a transaction is verified, everyone handling that transaction carries less risk. It is part of why we have been able to expand to multi-carrier shipping and bring payment options like Apple Pay and Google Pay into cross-border recommerce.

For the platforms themselves, trust is not a compliance cost, but an economic multiplier. Verified inventory converts better, disputes and returns fall, and higher-value categories become viable to list in the first place. In our case, luxury now makes up more than a quarter of our annual GMV, and that comes down to buyers trusting what they see.

How can technologies such as material analysis, item-level data, and authentication records make luxury verification more accessible beyond specialist labs or high-end offline channels?

Scientific authentication should not be confined to laboratories. It should travel with the item. Until recently, scientific-grade authentication required research instruments in a laboratory. While accurate, these were largely inaccessible to most of the market. However, at VivaTech this year, our deep-tech subsidiary InsightviewTech pre-launched a portable device that compresses that capability into handheld form. This enables optical sensing to analyze a material’s properties without any damage or sample preparation. The proprietary AI processes the data against our authentication database, and lab-grade results are returned in seconds on a mobile app.

The breakthrough is not only portability. It is interoperability.

Bunjang has also worked on areas such as international payments, shipping options, and lower cross-border costs. How do these buyer-experience improvements connect with the broader trust challenge in recommerce?

Trust does not end when authenticity is confirmed. That is when the transaction actually begins. Beyond authentication, an overseas buyer will consider the trustworthiness of an international C2C platform in any cross-border recommerce transaction. For example, how much will shipping actually cost, and will the item arrive? A buyer in Singapore or Jakarta who hesitates at checkout because they cannot estimate the total landed cost, or because their preferred payment method is not accepted, is experiencing a trust gap.

To mitigate this, we upgraded our cross-border purchase experience, enabling international buyers to choose between three shipping carriers with an upfront cost estimator to compare before checkout. This replaced a single-logistics-partner model where flexibility and predictability were limited. We also added Apple Pay and Google Pay, a first among Korean recommerce platforms, so international users can pay the way they are already familiar with. This improved our payment processes while delivering savings of at least ₩3,000, or USD 2, per item, roughly 5 to 10 percent.

Authentication builds confidence in the product. Buyer experience builds confidence in the platform. You need both.

The inclusion of secondhand data in music-related charts points to a wider shift in how resale activity is being recognized. What does this say about the mainstreaming of recommerce, especially in culture-driven categories such as K-pop merchandise and fashion?

The moment resale data enters official charts, recommerce stops being secondary.

A recent example is Bunjang’s partnership with Hanteo Chart, Korea’s first and most trusted music chart and a key data source for major Korean music shows and year-end awards. From July, its new Trend Chart officially incorporates C2C secondhand trading data from Bunjang, including overseas fandom activity on Bunjang Global, alongside B2C new-release data from Hanteo. This is the first time the chart industry has combined the two. The rationale reflects something we have long observed: album sales, streaming, and social activity were only ever capturing part of the fandom story. Fans engage by collecting photocards, trading merchandise, and buying memorabilia long after an album’s release, activity that was always happening but never officially tracked until now.

Resale is no longer a shadow market. It is becoming part of how culture itself is measured.

Looking ahead, what will determine whether cross-border recommerce in Asia can scale sustainably: better authentication, better logistics, stronger enforcement, better buyer experience, or closer collaboration between platforms and regulators?

Cross-border recommerce does not have a logistics problem. It has a trust problem.

Counterfeiting is ultimately a cross-border problem, and no platform, brand, or agency can solve it alone. That is why we would expect the platforms that thrive to be those that treat regulators as partners in market trust rather than guarding capability as an advantage. The winners will not simply move products across borders. They will move trust across borders.

Kaveh Wong of Xsolla Curine Academy on human creativity in the age of AI [Q&A]