Agilitas’ AI platform speeds up food, beverage R&D

An AI chip on a motherboard.
An AI chip on a motherboard. (Source: Getty Images/ Vertigo3d)

Silicon Valley AI-based startup Agilitas is helping food and beverage makers speed up research and development (R&D) and release products faster.

Founded earlier this year, Agilitas offers a suite of AI-based tools designed to assist with product formulation, reformulations and market analysis, tech executive veteran and founder of the company Akash Agarwal told FoodNavigator-USA in an exclusive interview ahead of a press release.

Agilitas' platform is built on an ever-expanding ingredient database and capabilities from AI technologies like OpenAI, Claude and Llama, the company explained. Additionally, Agilitas categorizes foods and beverages with help from its in-house food and data scientists, who can also reverse engineer food and beverage formulations of existing products, Agarwal said.

Agilitas’ goal is to capture what is inside “food scientists' heads,” so an organization can widely use that information to inform formulations, Agarwal added.

“Food formulation and food scientists were working in the Dark Ages, and I do not mean that with any disrespect to the profession or them. It is just that they got used to what was given to them, which was simple spreadsheets,” Agarwal said. “They could store a formula. They could define a formula, the ingredients, the compositions, and they could tweak it.”

He continued, “What they could not really do is what I call the ‘what-if’ analysis and the world changed with AI. This is a perfect layup for an AI use case because it is a hard problem, the number of ingredients is large, and it is getting larger.”

Agilitas use cases: Creating clean-label lemonade, muffins

Agilitas is currently available through a subscription-based service and already attracted paid customers, including an undisclosed leading lemonade brand and a better-for-you cereal company.

Additionally, Agilitas is working on a roadmap of new features, Thomas Hayes, co-founder and head of business development, told FoodNavigator-USA.

“The company will continue to build out capabilities that serve the needs of food R&D teams end-to-end. This spans from product benchmarking to formula development to nutritional analysis and more,” Hayes elaborated.

Through its capabilities, Agilitas provided the lemonade brand with insight to create a clean-label version of its product, including what a base formulation would look like and analysis of competitor products, Agarwal said.

The better-for-you cereal brand leveraged Agilitas to explore how to expand into muffins, using its base ingredients as a reference point, he added.

Agilitas’ technology not only helps figure out the way to formulate a new product, but it “extends across the gamut of use cases” for the food and beverage industry, including finding new ingredient suppliers, Agarwal said.

“Some of our customers are changing suppliers. The customer wants to launch the product in a new market and their existing supplier is saying, ‘You need to buy a minimum quantity.’ And they do not have the budget right now in this new market to buy that so they have to find an alternative supply for an existing ingredient and make sure that the product can still have the same capabilities,” he elaborated.

Will food scientists ditch spreadsheets for AI?

In 2024, food and beverage companies went from exploring AI use cases to implementing the technology into their workflows and investing in upgrades.

For instance, dairy producer Danone is working with Microsoft to incorporate AI into their system to boost operational efficiency and innovation.

Additionally, Google found in a survey of 376 senior executives at CPG companies and retailers that 57% of companies increased annual revenues by 6-10% with AI investments, while 30% grew revenue by more than 10% and 13% by 1-5%.

However, implementing AI into food and beverage workflows and processes might require some additional work as food and beverage professionals might be resistance to change.

In a survey of more than 1,300 companies, McKinsey found that 33% of generative AI (gen AI) high performers — the percentage that are seeing meaningful earnings from gen AI deployments — and 38% of all other respondents cited tech adoption and scaling as a challenge to capitalizing on generative AI efforts.

To address tech adoption, Agilitas worked alongside food and beverage scientists, using their insight to help build the platform, Hayes said.

“We were very deliberate in terms of developing the user interface and experience, and we worked hand-in-hand with food scientists, research chefs and product developers to build everything. We did not write a line of code until someone told us, ‘Okay, this is something that I would want to use, and here is how,’” Hayes elaborated.