Replicating coffee ‘cupping’ with AI: Start-up tracks coffee quality from crop to cup

Coffee tasting, otherwise known as coffee ‘cupping’, is the process by which the characteristics of a coffee bean are evaluated. Not dissimilar to a conventional sensory panel, when cupping, experts check a particular bean’s fragrance, aroma, acidity, body, flavour, and finish. However, according to ag and food tech start-up Demetria, this manual method is expensive and time consuming. Further, it is predominantly carried out by industry experts outside of origin countries, which can mean roasters and traders have ‘minimal visibility’ into the quality of the beans they purchase until they are exported. Demetria has also raised concerns that cupping is inaccessible to a majority of coffee farmers – notably the 12.5m smallholders who produce 60% of the world’s coffee beans. Consequently, growers are unable to determine or manage the quality of their crop. “Farmers’ inability to guarantee a consistent level of quality of bean means that a significant number of them receive just a base commodity price for their produce,” noted the start-up. “Only later in the process, when beans pass through an extensive and complex supply chain that incorporates traders, processors and exporters, is quality determine. [As a result], farmers receive on average just 2.5% ($0.7) from a cut of coffee retailing at $2.80.” Digitizing coffee aroma and taste Colombian-headquartered Demetria, which has its R&D capabilities in Israel, claims to have the solution. The start-up is digitizing ‘the most important quality variables of the coffee bean’: coffee aroma and taste. Demetria is leveraging near infra-red (NIR) sensors to analyse and ‘fingerprint’ green coffee beans for biochemical markers. Its artificial intelligence (AI) based platform, known as an ‘e-Palate’, then matches each bean profile according to the industry standard coffee flavour wheel, the start-up explained. For the ‘first time’, Demetria claims, quality and taste can now be assessed at any stage of the coffee production and distribution process. “The ability to discover the quality of green coffee beans is a game changer for an entire industry that’s relied on a primitive supply chain and artisanal processes for 300 years,” said Demetria co-founder and CEO Felipe Ayerbe. “It’s hard to believe that the world’s biggest roasters have effectively been buying beans with very limited knowledge about their quality, and that the majority of coffee farmers, the most critical players in the supply chain, don’t understand the quality of their own crops and hence are paid unfairly, threatening the sustainability of this $450bn industry. “Our technology delivers vital intelligence to ensure crop consistency and quality control, resulting in readdressing the economics of the coffee value chain to benefit every key player.” Demetria is currently focused on coffee, but said its technology can also be applied to other commodities, such as cocoa and sugar cane. Traceability on the beans, rather than on the pack Demetria is unique, a company spokesperson told FoodNavigator, because it is the first to deliver ‘true transparency’ of coffee based on the actual beans – rather than on the packaging. “We’re in discussions with major roasters who are extremely keen to deploy our technology and the ability to communicate traceability and quality to consumers is a major benefit to them.” ‘Everyone will win’, the spokesperson continued. “We have ambitious plans, we have the chance to be the start-up that revolutionises the entire supply chain for coffee, to help traders, growers, farmers, roasters and everyone in the coffee industry to move and improve their business models, with more coffee at less cost. “Everyone will win through the use of this technology, and coffee lovers have a lot to look forward to.” Forging industry partnerships Demetria has proven its technology works at scale, having completed a pilot with Carcafe, the Colombian coffee division of agricultural commodity traders Colcafe/ED&F Man. “Carcafe has identified a distinct high value ‘cupping’ profile and Demetria has built a matchmaking solution to determine which green beans will suit,” noted the start-up. According to Carcafe General Manager Sebastian Pinzón, Demetria’s technology has been able to replace many of its manual processes. “Carcafe worked with Demetria to develop a pilot application that allows anyone, anywhere, with a portable sensor, to detect a specific and high value sensory ‘taste’ profile of green coffee,” said Pinzón. “We are in the process of adopting and rolling out this ground-breaking technology to greatly increase the efficiency and effectiveness of our prospecting process at our purchase points and in the field. Demetria replaces many manual processes to determine vital data that’s previously been unavailable in the industry and is proving invaluable in making better decisions.” Demetria has also forged a working relationship with the Colombian National Federation for Coffee Growers, Federación Nacional de Cafeteros (FNC). Together, they are developing a series of smartphone apps designed to help farmers and their transaction points enhance traceability in the supply chain. Ultimately, the goal is to better control and track bean quality, to ensure it is priced appropriately. FNC President Roberto Velez said it is ‘extremely important’ that Colombia continues to strengthen its position within the specialty coffee market. “Through the use and affordable democratization of sensors, cloud computing and other technologies, coffee growers will be able to manage each stage of the production process in a simple, timely and accurate manner, and more importantly, evaluate the quality of their own crop and gain a price that reflects this. Technologies like Demetria empower the FNC to achieve its goals and play a key role in the sustainability of our industry.”

Replicating coffee ‘cupping’ with AI: Start-up tracks coffee quality from crop to cup

Ag and food tech Demetria says it is pioneering the automated analysis of coffee bean taste using artificial intelligence and data cloud technology.

Related Topics