How generative AI could boost plant-based protein functionality

AI is shaping just about everything. In the world of food and beverage, for example, it is allowing companies to get more extensive and up-to-date consumer insights​, optimise operational efficiency​, and even recognise cows​.

AI has not left the plant-based world untouched. Paris-based start-up AI Bobby uses generative AI to help design and optimise proteins for use in plant-based meat and dairy analogues.

How AI can boost R&D​Much like in the pharma space, AI can boost innovation in food and beverage. “It’s able to, from a large quantity of information, give the most probable answers,” AI Bobby founder and CEO Dominik Grabinski told FoodNavigator.

It can increase the speed of finding a solution by a factor of four to 10, he told us, and the likelihood of finding a solution from between 20% to 50%.

AI Bobby aims to help plant-based protein producers improve the functionality of their proteins in a more efficient and accurate way using generative AI.

How can plant-based protein be improved?​Certain plant-based companies, such as Beyond Meat, have been struggling in recent months. Some fear that consumers are turning away​ from the plant-based category.

The biggest reasons for these things, Grabinski told us, is protein functionality and fat functionality. “The functionalities of these proteins – it can be in terms of gelling, it can be nutritional, it can be mouthfeel – are not there, especially if you compare it to the proteins that they need to replace, which are the animal proteins.”

AI Bobby wants to fix these problems. It “understood that the current stage of knowledge and science that has been built is not able to move fast enough in order to build it,” Grabinski told us. So the company opted to use generative AI. 

With generative AI, research into the functionality of the protein can be sped up and made more efficient.

Which areas of plant-based is AI Bobby focused on?​In order to get the most out of AI, it must be an expert in the domain in which it operates. “We need to train it domain by domain by domain and go very, very deep.”

The domain which the start-up is tackling first is gelation. Gelation, Grabinski told us, is important for the mouthfeel and texture, as well as the ‘easiness of formulation.’

Improving this aspect of the protein could increase its functionality, he suggested, which may reduce the need for additional ingredients within a plant-based meat analogue to fulfil these same functions. This, in turn, has the potential to put the cost of the product down.

Function to structure ​In the future, the company aims to be able to link function to structure. This essentially means that they will be able to help plant-based producers using precision fermentation to better design the protein that they craft inside of seeds or cells in order to suit the functionality necessary for specific foods.

“Nobody told you that casein or whey protein produced by precision fermentation should be equivalent to what the cow was producing. We can do it better,” Grabinski told us.

Eventually, he suggested, he would like his customers to be able to design proteins that fulfil the key functionalities in a range of animal-based proteins like collagen, gelatin and egg white, but within plant-based sources.