‘How sweet is sweet?’ AI-powered algorithm digitises taste for food and beverage makers

When veteran entrepreneur Yuval Klein was just six years old, he tasted a drink that would change the course of his life.

‘Rosetta’ is a sweet and bitter Israeli beverage made from sugar and almonds. When served cold on a sweltering hot day, as it was on this occasion in Jerusalem, 1967, it is ‘the most amazing drink ever made’, recalled Klein.

The entrepreneur has long been trying to recreate this beverage with the ‘right’ balance between bitterness and sweetness. But after an extensive career (Klein has founded or co-founded at least six companies in his lifetime, including sugar reduction start-up BlueTree Technologies​), he has recently found more time to dedicate to this project.

Around eight years ago, the entrepreneur set about studying different sugars and their sweetness to help develop the perfect Rosetta recipe. To help the process, he developed an app designed to describe sweetness.

Realising that food and beverage makers could also benefit from understanding what ‘sweetness’ means for consumers, Klein founded MAMAY Technologies in 2022.

As founder and CEO, Klein led the development of an AI-powered algorithm to determine the ‘objective’ sweetness of product. Taste, he told FoodNavigator, can be digitised.

Starting with the digitisation of sweetness, the start-up uses laboratory equipment to conduct physical and chemical testing on food or beverage products. The tests analyse these products’ molecules.

Understanding the impact of these molecules – and their impact on taste and smell when combined – allows MAMAY to objectively map their sweetness profile.

The company recently launched its first software as a service (SaaS) platform designed to provide an ‘objective simulator of product taste, odour, and feel’. The platform can quantify the impact of sweetness based on more than 70 eligible sweeteners in food and beverages, we were told.

The technology is powered by artificial intelligence to manage the vast amounts of data required to digitise taste.

As to whether the platform could end up replacing sensory panels, the founder suggested it could replace at least some of them. Sensory panels would still be required to determine whether consumers like or dislike a product, but MAMAY’s AI-powered algorithm could do the rest.

The start-up has already partnered with businesses working in water, coffee, milk, and tea categories, who can use the platform to determine the taste of their own product, as well as that of their competitors.

And sweetness is ‘just the beginning’, explained MAMAY’s founder and CEO. In the coming months and wants to offer services for saltiness, bitterness, sourness, and spiciness.