As counterfeit detection becomes common, reporting-evidence services open up

AI can find a listing, but a person still has to report it

On November 4, 2026, the Ministry of Agriculture, Food and Rural Affairs (농림축산식품부) and the Korea Agro-Fisheries & Food Trade Corporation (한국농수산식품유통공사), a public body that supports food trade, announced a programme to monitor and block counterfeit K-food products. Applications run from November 2 to 16, and five K-food exporters will receive support. With only five companies selected, this looks less like a broad rollout and more like focused support for testing the real reporting process.

The covered marketplaces are Taobao, Tmall, 1688, Lazada, Alibaba.com, and AliExpress. Each company can receive support for one brand, covering products produced, manufactured, or processed in Korea. The programme supports monitoring and takedown reports across major online platforms in China and Southeast Asia.

AI monitors suspected infringing products and classifies the type of infringement. Finding a suspicious listing does not automatically remove it: the participating company must confirm that it is an actual infringement and approve the report. The marketplace reviews the evidence, then may remove the listing, ask for additional materials, or leave it in place.

Reports need grounds such as trademark rights, which protect a brand name and mark; design rights, which protect packaging shape; or copyright, which protects photographs and text. The rights holder's name must also match the applicant company. In the selection review, companies that submit three or more suspected listing URLs receive higher scores. In other words, organizing already-found URLs and rights documents matters as much as technology that finds listings well.

The programme runs from selection until December 24. The Intellectual Property Office (지식재산처), the government body responsible for intellectual property policy, has also said it will increase the blocking of counterfeit K-branded products overseas to 300,000 cases by 2027. That makes the time companies need to review candidates and approve evidence more likely to become scarce than detection volume itself.

The bottleneck is not the search box. It is the evidence drawer.

Consider a hypothetical case: Kim runs a sauce manufacturer with 12 employees. The company exports gochujang sauce to China and Southeast Asia, but has no dedicated counterfeit-product employee. A member of the overseas sales team handles orders and customs clearance, and responds only when a distributor sends a suspicious listing.

The employee pastes the listing URL into a document, saves screenshots, and searches the company shared folder for the trademark registration certificate and authentic packaging photos. They must compare the logo placement, volume label, and manufacturer address between the authentic and suspected products, then write the reason for the report in Chinese or English. If the same seller relists the product under a different product name, they first have to find the earlier reporting documents again.

An even harder issue is that not every cheaply sold product is counterfeit. Genuine goods may be resold outside official distribution channels, so price and seller name alone are not enough to support a report. If a listing uses authentic photos but ships a different product, records of a test purchase, shipping label, and unboxing photos may also be needed.

The new support programme can reduce the time spent finding suspicious listings and grouping similar infringement types. But Kim needs a screen that shows which of the daily candidates to approve. It needs to show the grounds for suspicion, the same seller's history, rights already held, missing evidence, and the likelihood that a report can be filed in one place.

The next need is a service that automatically drafts the different reporting forms required by each marketplace. Kim could upload a trademark registration certificate and authentic product photos once, then receive an AliExpress report form and Taobao comparison materials tailored to each platform. If a seller challenges the report, the service could place the materials from the original report next to the new claim and show only the items that need a response.

Some decisions still belong to people. The brand must decide whether a product is truly counterfeit, whether it is a resale by an official distributor, whether a test purchase is needed, or whether to move to legal action. The service should not make those decisions; it should eliminate repeated work finding and copying the same files.

Looked at another way, as detection technology spreads quickly, it becomes harder to charge for detection alone. The remaining work is turning findings into reportable evidence, meeting response deadlines, and remembering whether removed sellers reappear. For small K-food companies, the opportunity is not a bigger search engine but a tool that closes the work after detection from start to finish.

Abroad, services are bundling the steps after detection

Fila Manila, a Philippine food brand operating in the United States, sells banana ketchup, sauces, and other products on Amazon. The company used Amazon Brand Registry, which lets trademark-registered brands report infringing listings, together with Transparency, which adds a unique code to each product. Brand Registry and reporting are free, while unique codes are charged according to the number issued.

The company said imitation products and unauthorized sales had been costing it about 20% of its business, and that sales increased by 20% after adoption. The figures are the brand's own claims, published by Amazon, and need independent verification. What matters is that it connected reporting with a preventive method that verifies authentic goods by code before shipment.

In China, Alibaba's Intellectual Property Protection Platform (IPP), its intellectual-property reporting channel, accepts reports for Taobao, Tmall, 1688, AliExpress, and other platforms through one account. Rights holders register their identity and trademark documents, then submit infringing URLs and comparison materials; sellers can submit counterarguments. Earlier practical guidance for AliProtect and TaoProtect described basic registration and infringement reporting as free, while test purchases, authentication, and litigation costs were separate.

The Royal Canin counterfeit pet-food case did not end with listing removal. In 2016, Taobao made a test purchase of pet food priced at about KRW 17,000 and sent it to the brand. After the packaging and contents were found to be counterfeit, the case proceeded to a police investigation and civil litigation. In 2017, a court ordered the seller to pay Taobao about KRW 20 million in damages.

Food Huggers, a US food-storage brand, used Red Points after counterfeit products spread from Alibaba to other online marketplaces. The service sets the brand, countries, and marketplace coverage to protect, charges a flat fee, and manages suspected-listing detection and removal requests through a dashboard.

Red Points said it found 5,322 cases across more than 74 online marketplaces and removed 3,542 of them, with a 96.9% takedown success rate for cases that could be reported. These are figures published by the service provider and need independent verification. Still, what customers bought was not search results, but an operating model that repeatedly processed reports across multiple marketplaces.

Things you could build now

1. Reporting evidence pack builder

• What it does: Takes trademark registration certificates, authentic-product photos, and suspected-listing screenshots, then creates marketplace-specific report forms and comparison materials.
• Who uses it: A sauce or snack manufacturer where one overseas sales employee handles both exports and counterfeit reports.
• Why now: More suspicious listings are being found, while limited staff still have to repeatedly copy the same evidence into different places.
• First screen: Three columns for rights held by brand, reportable listings, and missing evidence.

2. Suspected-product approval inbox

• What it does: Collects candidates from multiple detection tools and sorts them by likelihood of being counterfeit, likelihood of being genuine resale, and whether a test purchase is needed.
• Who uses it: An employee at an export agency that manages overseas online sales for several food brands.
• Why now: As detection volume rises, a person cannot sustainably read every listing from the beginning.
• First screen: Shows authentic and suspected product photos side by side, with buttons for approve report, hold, and test purchase.

3. Reappearing-seller tracker

• What it does: Alerts a brand when a removed seller returns under a changed store name, product name, or image, linking the new listing to the earlier report record.
• Who uses it: A kimchi or ramen export-brand manager who completed a large takedown but keeps seeing the same products return.
• Why now: Support that removes one listing at a time cannot easily stop repeated sales through new accounts and URLs.
• First screen: Lists sellers who reappeared this week, their previous removal count, the images used, and linked stores.

4. Packaging authenticity-check page

• What it does: Assigns a different verification code to each product so consumers and local distributors can check authenticity by phone.
• Who uses it: A health-food or sauce manufacturer entering overseas online marketplaces for the first time, where distribution routes have become more complex.
• Why now: For food products, removing listings alone is not enough; a preventive tool is needed to connect the delivered product with production records.
• First screen: Includes a code-entry field, authentic-product photo, production date, official sellers, and a button to report a suspected product.

What to check today

Ask a person responsible for an exported food company to show you, from beginning to end, every file used for one recent report of a suspicious listing. If, within 30 minutes, the trademark documents, listing URL, authentic-product photos, and comparison materials are scattered across three or more places or cannot be found at once, an evidence-pack builder is worth testing first.

Why this matters where you are

Check whether marketplaces in your target countries require different evidence, forms, and response steps for infringement reports. The platforms, rights documents, and enforcement paths may differ from the examples here, but the operational work of collecting, reviewing, and reusing evidence remains visible. Start by mapping the files and decisions involved in one real report before building more detection.