NotCo is a plant-based food company with two faces: a consumer brand and an artificial intelligence platform for other food makers. The consumer side sells plant-based alternatives such as NotMilk, NotMayo, NotBurger, NotChicken and NotIcecream. The business side, NotCo AI, offers Giuseppe AI, which the company describes as a vertical AI model stack for innovation, formulation and growth.
Giuseppe AI for formulation
Giuseppe AI covers product development from idea to market in three stages. Concept generates and validates product ideas from consumer data and competitive intelligence and produces a brief for research teams. Discovery helps scientists choose formulas using a proprietary ingredient catalogue, domain-specific models and multi-agent AI. Elevate runs digital experiments to reach the final formula faster.
The platform models cost, regulatory and process constraints before trials begin, and aims to reduce physical trials. For a formulator, the practical meaning is a search tool over ingredient combinations, ranking candidates by predicted match to a target product. Predictions still need bench validation, because texture, flavour release and stability depend on processing details that models capture only partly.
NotCo reports figures for the platform, including 16 times faster development, 11 times higher research productivity and a full match to brief attributes. These are vendor claims, so ask what baseline they were measured against. A prospective customer should ask for case-study data, including the products studied, the time recorded before and after, and who measured it.
The site names more than 20 global consumer goods companies as users and lists PepsiCo, Magnum, Mondelez International, Barry Callebaut, Kraft Heinz, Ferrero, Grupo Bimbo and Arcor among customers. Published case studies cover snack innovation for PepsiCo, ice cream for Magnum and chocolate for Barry Callebaut, which shows the platform is used across several categories, not only plant-based foods.
Plant-based consumer products
The consumer range includes NotMilk, NotMayo, NotBurger, NotChicken, NotHotDog, NotChori, NotSquares and a protein bar. The company says its AI helps recreate the taste and texture of animal products using plants. Such products give a developer real market examples of plant-based formulation, and they show how ingredients such as pea protein are used in practice.
Plant protein ingredients feed many of these categories. Pea protein supplies body and protein in dairy alternatives and bars, and texturised proteins form the base of burgers and nuggets. Context on these materials is on the pages for pea protein and textured vegetable protein. Ingredient specifications for platform projects come from each customer’s own suppliers.
The website mentions a food service and customers section and a B Corporation page. A food service team supplies products to restaurants and institutional kitchens, and the B Corporation reference points to a certification that the company should document with its score and date. Buyers who plan to cite either point should check the primary pages.
The company also operates an online store and a recipe section. For a developer, the recipe pages show how the products are meant to be used, and the store provides a way to buy retail packs for benchmarking. Benchmarking against market products is a standard step in sensory work and label review.
How a buyer would evaluate the offer
The first question for the AI platform is data. Ask what data the models train on, how customer data is separated, and who owns the formulas that result. The site says it holds a decade of normalised data from instrumented protocols and production programmes, and it references a trust centre for security compliance. Review that documentation with the legal team.
Next, test on a real project. Choose a product with a clear target, such as a plant-based cream cheese with specific texture and cost limits, and run the platform in parallel with conventional bench development. Compare the number of trials, time to a passing prototype and the performance of the final product on sensory and shelf-life tests.
Ingredient access matters too. The platform uses a proprietary ingredient catalogue, so check whether it can work with the customer’s approved suppliers and constraints, such as allergen policy, label limits and regional availability. A formula that depends on an ingredient the customer cannot buy at scale, or at the right price, has little value.
Commercial terms should cover pricing model, term, exclusivity, intellectual property in formulas and exit rights. A quote is required to see pricing. Evaluate total cost, including internal staff time for validation, against savings in trials and speed. A small pilot with defined success criteria limits exposure before a larger contract.