KANSAS CITY — Think of artificial intelligence (AI) as potentially a way to accelerate innovation in the food industry.

Ingredient selection is one example.

“Artificial intelligence has the potential to make a significant impact on how we identify and select ingredients for new product development,” said Ashley Robertson, director, global marketing and communications for Corbion and based in Lenexa, Kan. “By leveraging large datasets, AI can predict how different ingredients interact with one another and how they might impact the final product’s taste, texture and nutritional value. This approach can lead to innovative ingredient combinations that may not otherwise have been considered, ultimately enhancing product innovation and competitiveness.”

AI also could minimize the traditional trial-and-error approach in product development, she said.

“By using simulation and predictive analysis, AI can expedite the formulation process,” Robertson said. “This would enable faster iterations based on feedback, which could streamline the journey from concept to market.”

The food industry already has several AI examples.

Months, not years

Brightseed, San Francisco, uses Forager, its computational platform driven by AI, to identify bioactives in nature and their health benefits. Through Forager, Brightseed in three months identified bioactive compounds out of hemp hulls that help maintain a healthy gut and metabolism, said Jan-Willem van Klinken, MD, senior vice president, medical, scientific and regulatory affairs. He added the industry standard is three years. Those compounds, N-trans caffeoyltyramine (NCT) and N-trans feruloyltramine (NFT), led to the creation of Brightseed’s first product, Bio Gut Fiber.

Brightseed partnered with Manitoba Harvest, Winnipeg, Man., in February to launch Manitoba Harvest bioactive fiber with Bio Gut Fiber. The ingredient has six grams of fiber and is available at Whole Foods Market and through Amazon.

“Forager is built on machine learning and works in three parts: it predicts compounds from plants, fungi and bacteria sources; it predicts their activity in the body; and it predicts which source contains each specific health-benefiting bioactive,” van Klinken said. “Brightseed’s scientists source plants from around the world to feed into Forager, and we train our AI models with highly curated biomedical insights to generate the world’s largest library of natural compounds and their predicted impact on human health. Forager is able to predict with unprecedented speed and accuracy which bioactive compound impacts a specific health benefit in the human body.”

Forager allows Brightseed to reduce the time needed for bioactive discovery, biological validation and ingredient formulation from years to months, he said.

Rapid bioactive profiling gives Brightseed’s partners insights into phytochemical and bioactive counts as well as predicted health benefits in six to eight weeks, van Klinken said.

“Rapid bioactive profiling can then be followed up by deeper bioactives profiling in vitro, in vivo and clinical validation depending on a partner’s commercialization goals,” van Klinken said. “We also work with partners in targeted health areas within the microbiome.”

ADM, Chicago, is partnering with Brightseed to use Forager to examine the chemical interactions between dietary plants and gut microbiomes. Forager identifies bioactive compounds in plants and provides insights about cellular changes. The partnership aims to uncover bioactives created exclusively through interactions with gut microbes such as salicylic acid. The companies hope to launch functional ingredients in foods, beverages, dietary supplements and medical foods by 2025.

“We have profiled thousands of plants and have the world’s largest database of natural compounds mapped to their predicted mechanisms of action,” van Klinken said. “On average, Brightseed adds over 130 plants to our library a month to our compound libraries, which is a rate of about 6 plants a day. Our goal is to map the entire edible and medicinal plant kingdom, which amounts to roughly 80,000 species.”

Brightseed has mapped 23 consumer health areas as well. The company develops bioactives for partners in dietary supplements, foods and beverages, specialty nutrition, and medical foods.

“Our current pipeline includes novel products for near-term development that address large, unmet consumer health needs in the areas of gut health and immunity, metabolic health, weight management, cognitive health, sleep, maternal health and other targeted health areas through microbiome,” van Klinken said.

Intelligent ingredients

Neil Foster, head of strategic partnerships at Dublin-based Nuritas, explained how his company benefits from AI in a SupplySide West session Oct. 24, 2023, in Las Vegas.

“We are using artificial intelligence to create intelligent new ingredients,” he said.

Nuritas has collected data for eight years, especially data for peptides, which are small proteins made up of amino acids. Nuritas has over 6.2 million peptides in its library.

“Every single living creature, plant, bacteria is effectively like an encyclopedia of data and information,” Foster said.

He then gave a library comparison. If researchers wanted to find a peptide with a specific biological benefit, like managing carbohydrates consumed or increasing muscle protein synthesis, they could read every single “book” in the peptide library.

Or they could turn to AI.

“So what AI enables us to do is read these books infinitely quicker, have perfect recall and perfect memory, and know which word is found in which place, and to speed up the whole process,” Foster said. “AI enables us to find those (peptides) far, far, far faster. It also enables us to predict the problems we might have to not just finding the molecule but to know it’s something we can access and is going to be usable and is going to be affordable and is going to be sustainable.”

Through AI, Nuritas created PeptiStrong, a fava bean ingredient shown to boost muscle protein synthesis, reduce muscle protein breakdown and decrease exercise-induced inflammation. PEP, Solon, Ohio, launched PEP bars containing PeptiStrong this year.

In supply chain management, AI potentially could generate more accurate forecasts, identify supplier risks, recognize unusual behavioral patterns, optimize pricing and transportation plans, determine inventory levels, and provide financial analysis, according to Info-Tech Research Group, a technology research and advisory firm serving over 30,000 information technology (IT) professionals. Adoption barriers are selecting the right vendor and implementing strategies.

“AI could significantly enhance the sourcing of ingredients by making supply chain operations more efficient and sustainable,” Robertson of Corbion said. “Through the analysis of global market trends, environmental factors and logistical data, AI has the potential to pinpoint optimal suppliers for sustainable and healthy ingredients. Furthermore, AI might refine inventory management and demand forecasting, contributing to more responsible and sustainable sourcing methods.”

She added, “AI might also influence the food ingredients category by automating quality control, predicting food trends, optimizing nutritional content and assessing the environmental impact of production processes.

“These applications could lead to higher quality products, more proactive trend adoption, healthier food options and more environmentally friendly practices, showcasing the broad potential of AI to benefit the industry and consumers alike.”