Osmo Smells Opportunity with Scent-based ID

Published: February 3, 2025
  • Computers are using Osmo’s sensors and AI to sniff out fake products during authentication of high value sneakers.
  • In the long term, the system could also leverage scent markers for specific products that are undiscernible to the human nose, but that can be detected by the Osmo solution.

Identifying unique objects in a supply chain often relies on barcodes, RFID or other optical technology. When it comes to detecting a counterfeit product, these technologies provide an indication there’s a problem, but not always.

In fact, product authentication can also be detected through other senses without requiring a unique identifier: smell.

That’s the premise of Osmo, a digital olfaction company, where officials are rethinking how to authenticate products in supply chains and stores with AI-powered scent sensing. The technology leverages the fact that every type of product has a specific scent, based on where it was made, its materials and where it’s been.

The company already has its background in creating scents for perfumes. And although the majority of its business focuses on designing fragrance, scent detection has become a key business strategy for Osmo,.

“Our mission is to give computers a sense of smell,” said Alex Wiltschko, Osmo’s CEO.

Two Years of R&D

The New York City-based company’s engineers have been developing a sensor-based system to detect and analyze the chemical signature of a product. Several years ago, it achieved a $60 million round of funding led by Lux Capitol and Google Ventures.

Development centered around sensors, as well as machine learning, data science, psychophysics olfactory neuroscience, electrical engineering and chemistry. The goal was to detect aroma molecules specific to products. Work included using graph neural networks (GNNs), designed to predict the odor descriptors for individual molecules, without requiring human rules.

“We’ve solved that after two years of very intensive R&D—it works in the lab and now there’s a piece of it that works now in the field,” he said.

Identifying High Value Sneakers

Collectors often buy high value goods directly from the brand, or through a reseller that offers authentication services to ensure the product is not a fake. One example is high end sneakers, a product that Osmo chose to target first. A new sneaker, manufactured by a specific brand and factory, will have its own identifying odor. While counterfeiters can replicate the products, they can’t replicate the smell.

“They can look the same but they can smell totally different,” Wiltschko said.

Sneaker seller Stock X verifies and sells luxury goods including high-end sneakers, accessories and clothing. For each product, Stock X provides a stamp of authenticity for the customers so that they can be assured the pair of shoes or other collectible is genuine. It is now employing Osmo to automate some authentications.

20 Second Analysis

With the solution, inspectors insert a small sensor probe—connected to a computer device running the Osmo software—into a shoe box. The process takes about 20 seconds as the system analyzes the olfactory readings. The technology was first put to use at Stock X in late 2024.

With AI, Osmo has trained the software to differentiate between real and fake versions of many different SKUs or sneakers that are problematic, “so it trains the system over time,” Wiltschko said.

As scent molecules are detected, the system bombards those molecules with energy and measures the ways the molecules respond. It uses machine learning to update data about what characteristics are common for a fake product, or a new authentic product.

Saving Time, Boosting Accuracy

For personnel who work at the company as verifiers, the system is considerably faster than a visual check of the shoes or other products, Wiltschko said. In the manual method, he argued, they spend more than 20 seconds per shoe verifying.

The extended ambitious goal for Osmo officials is to deploy the technology for a wide variety of products.

“In the long term, the system should be operating at all U.S, ports,” Wiltschko said. “The problem of counterfeits here in the U.S. should be stopped at the borders” he added that while sniffer dogs could potentially be trained to track for counterfeits, that would be too expensive.

The Customs and Border Patrol indicated that over 90 percent of all counterfeit seizures occur in the international mail and express environments, which are channels that small, e-commerce packages destined for the U.S.

In fact, potentially thousands or even millions of people are walking around in high-cost brand sneakers that they have no idea are not genuine, Wiltschko pointed out.

Beyond Sneakers

Authenticating sneakers was a great test case for the technology since they are an evocative challenge for brands and buyers, and each shoe has a complex blend of molecules.

The technology could be used for products such as wines and liquors or perfumes, according to Osmo officials. In some cases, users could open a bottle and measure odor related to a single product on a pallet, (like one wine bottle in a case).

Osmo is working with other companies regarding authentication of other products. “We’re always looking for more interesting partners to help us advance the technology and the business,” Wiltschko added. “We’re looking at folks who have more sensitive supply chain applications,” which includes logistics companies, brands and buyers.

It could be a feasible alternative to technology like RFID which provides a unique identity to a single product as RFID tags can be cloned. Odor creates a snapshot of a product’s history as well, Wiltschko said—how a product was manufactured and how it was stored and shipped, could potentially by identified by the odor.

Adding Tags

Beyond sneakers, the solution could work with other products that may not come with an inherent scent. Osmo is investigating providing its own marker tags with a distinct odor that serve as olfactory watermark. The company is developing a new product that acts as an invisible chemical markers into products that can be detected by Osmo sensors later.

The marker emits only a faint smell of oxidizing compounds. “We’re really good at making things that smell great but we also can create things that smell like nothing to people, but are very, very apparent to our sensors,” said Wiltschko.

Additionally, the company is looking at health and wellness applications to detect the molecules that humans produce, when they are well, but also when they are not. The same technology could be used to measure athletic performance.

Learn More: