ProfilePrint, an AI company that reads the molecular fingerprint of food ingredients to judge their quality, has added a matching tool that works straight out of the box, bringing its technology closer to farmers, AgNavigator reported. Until now, clients had to collect samples and train and validate a model for each product, a process founder and chief executive Alan Lai said took a few months before they saw any value.
The new approach skips that step. A client picks a target ingredient, and the system filters and matches other lots by how closely their molecular fingerprint matches it. Lai said this is possible because of the database of ingredient fingerprints the company has built up over years of commercial use.
That moves the tool upstream, so exporters and agricultural cooperatives can test their produce against the same quality targets that ingredient buyers and manufacturers use. "You can produce a lot of yield, but if the volume is creating low-cost products that no one is going to pay a higher price for, then you are creating excess surplus that is going to cause the price to fall," Lai said. The company is working more with cooperatives and producer groups that want to see how farming practices affect crop quality and market value.
ProfilePrint started in coffee and tea and now covers cocoa, grains, rice, juices, fats, herbs and spices. Because everyone in a supply chain can work from the same target profile, traders and importers can check that a lot matches their customer's needs before buying, and in some cases hold less stock.
Lai said the company first had to learn what large food and drink makers want before it could take that knowledge back to farmers, to help them improve yields and farm economics.
Reporting by AgNavigator.
Photo: Fernando Rebêlo / Wikimedia Commons (CC BY-SA 3.0)
Source: AgNavigator





Comments
(0)