The Center for Data Innovation spoke with Vidyuth Srinivasan, co-founder and CEO of Entrupy, a New York-based firm that uses artificial intelligence (AI) solutions to authenticate consumer goods. Srinivasan explained how each new authentication strengthens the AI model’s accuracy and discussed the future of product authentication beyond luxury goods and in the metaverse.
Becca Trate: What problems did you identify that led to the founding of Entrupy?
Vidyuth Srinivasan: At its core, Entrupy provides businesses, governments, and other organizations with solutions that safeguard them and their customers from the negative impact of product counterfeiting, an enormous problem that causes a massive drain on the global economy. Not only do counterfeits displace billions of dollars in legitimate business activity, subsequently reducing tax revenues for public services, investment, and innovation, their illicit trade funds additional criminal activity. Furthermore, counterfeiters often rely on forced labor to manufacture goods while also using materials and methods that expose consumers to potentially dangerous products and toxic chemicals.
Our flagship solution, Luxury Authentication, was created to address one of the most frequently-targeted categories for counterfeiters: high-end handbags and accessories. These products are highly-desirable, and much of their exceptional value comes from their exclusivity. Produced in limited quantities, demand exceeds supply, and bad actors exploit the difference for profit. Using Entrupy’s AI-driven technology gives businesses trading in these luxury items a fast, user-friendly, completely objective, and infinitely scalable method to verify whether items in this category are authentic. This in turn provides trust guarantees that consumers didn’t have when they stumbled upon a coveted piece at a small shop, for instance.
Based on similar technology and approaches, we’ve also created sneaker authentication technology, addressing the same problems in another fast-growing segment that is an easy target for counterfeiters.
At a higher level, we’re plugging the trust deficit in commerce transactions that are increasingly impersonal and have unstructured supply chains, and that’s imperative to the growth of resale, which is one of the fastest growing segments of the retail economy.
Since launching Luxury Authentication, we’ve continued to develop solutions targeting other issues and product categories. For example, our fingerprinting solution enables item-level identification. This has proven incredibly effective at preventing return fraud and it also has great utility for tracing an item’s provenance.
Trate: How does AI authentication work, and how does it compare to other forms of authentication?
Srinivasan: Entrupy’s authentication solutions use a combination of AI and microscopy to objectively assess the authenticity of an item. When the item is scanned using our proprietary technology, a series of microscopic images are collected and analyzed by machine learning algorithms. The algorithms are trained with millions of microscopic images from known-authentic and known-counterfeit products. Based on the results of this comparison, the AI will either verify the item’s authenticity or return an “unidentified” result. Each scan from an Entrupy device becomes part of the database, further training the algorithms and making our solution smarter and more accurate.
Our solution has a 99.1 percent accuracy rate. While it is highly unlikely that an inauthentic item will be misidentified as authentic, we provide a financial guarantee to protect the certificate holder from financial losses if the technology does provide errant results. This is critical because it helps the ecosystem preserve the level of trust in the ecosystem.
Trate: Entrupy also detects “switch and return” fraud. What is this, and what AI solutions protect consumers and retailers?
Srinivasan: “Switch and return” fraud is a particularly insidious form of retail crime that involves purchasing one product, switching one item for an inferior or counterfeit item, and then returning—and obtaining a refund for—the switched item. This could be buying an authentic designer bag and sending back a fake, or it could be switching an old or broken iPhone for a shiny new one. Either way, the thief keeps the money and the item, and it’s possible that the swapped item will remain in the supply chain.
What many people do not know is that every item, even those that are mass-produced, has unique characteristics that make it different from other objects. For example, two iPhones that are the same model and are manufactured in the same factory on the same day, will still have minute differences in characteristics like the physical textures. Entrupy Fingerprinting leverages microscopy and artificial intelligence to capture and record these differences, enabling that item to be re-identified at any point in its lifecycle.
To prevent return fraud, retailers simply “register” high-value items by using the Entrupy app to capture a single microscopic image of the product. The image, stored securely in the Entrupy Cloud, acts as a digital fingerprint. At any point in the product lifecycle including at the time of return, that exact item’s identity can be verified.
For example, if a retailer sells a high-end handbag that has been registered using Fingerprinting, then that item is returned, the store or warehouse associate can verify that it is, in fact, the exact same item. The associate need only capture a new microscopic image of the product, which is then compared to the registered fingerprint using Entrupy’s patented artificial intelligence technology. In seconds, the associate will receive the results. If it is a “match,” they can proceed with the return. A “no match” result means the item is not the same and the associate should not accept the return.
Trate: What is digital fingerprinting?
Srinivasan: Entrupy Fingerprinting is an unobtrusive system of product authentication that has two significant advantages over traditional identity solutions. First, unlike barcodes, radio frequency identification (RFID) tags, holograms, and QR codes that require something to be attached to or printed on the physical item, fingerprinting is completely non-intrusive. The identity is recorded using just a single microscopic image. This makes the solution much faster and easier to implement at any point in the supply chain while enabling products to keep their integrity.
Second, fingerprinting is significantly more secure than traditional solutions, which generally lack security and are therefore more prone to manipulation. For example, there have been several instances where an existing QR code was replaced with a lookalike. When the QR code was scanned, users would be unknowingly redirected to malicious content or trojan scripts, revealing their valuable personal information to bad actors. This vulnerability to corruption creates enormous risks around data privacy. Fingerprinting as a technology is applicable to any physical product without requiring physical modification of the product. Categories for which Entrupy has developed significant support include designer bags and accessories, watches, beauty products, wine and spirits, electronics, and medical equipment.
Trate: Entrupy primarily authenticates luxury goods and sneakers in secondhand markets. How do you see the company expanding to authenticate other goods in traditional and e-commerce?
Srinivasan: Expanding the application of our authentication solution is really a matter of building a significant database of images for the comparison. We are certainly amenable to working with organizations to determine whether such a database can be created to successfully address their needs. It also depends on the segment of products where there is a clear trust deficit due to counterfeits, a lack of clear standards of authentication, or even a chain of custody.
We’ve also been working on a solution, Entrupy Brij, that addresses issues related to trade across decentralized platforms. It’s an app that lets anyone quickly and easily mint NFTs of physical objects to connect ownership in the real world with ownership in the evolving metaverse. Once the NFT has been minted, the owner can perform various valuable actions with it, like importing it into the metaverse, using it for proof of ownership, showing provenance, tracking asset value, verifying authenticity, and trading on non-traditional marketplaces and platforms.
We envision Brij as an open solution that will bridge the digital and physical divides. Through partnerships, integrations, and other collaborations, it could ultimately become a de facto standard across relevant blockchain ecosystems, wallets, marketplaces, manufacturers, and communities.