How AI Is Revolutionizing Product Development

Like most food scientists, Rachel Zemser is something of a novice when it comes to artificial intelligence (AI). But when the Almond Board of California asked her to play around with Tastewise IO and use the TasteGPT tool to come up with some product ideas around gut health and other dietary needs of Generation Z consumers, she dove in.

Zemser used her Tastewise findings to develop a few product ideas leveraging currently trending flavors and tastes, including a vegan aioli made with almond oil instead of eggs and an almond-chocolate candy bar. But the Almond Board’s favorite was an Almond Champurrado Mix (Mexican-style hot chocolate) based on super-fine almond powder from supplier Cache Creek Foods that the AI engine discovered.

“I used AI in combination with what’s trending, what I’ve seen myself in the world, and how to use food science with what’s trending and turn it into something with almonds that never had been done before,” says the principal of A La Carte Connections.

In important ways that meet the needs of legions of food product developers who are trying to work more effectively every day, AI is raining down from the digital cloud and inundating the industry with thousands of unforeseen ideas for how to make better or unprecedented ingredients, foods, and beverages. It is already helping companies save thousands of hours, millions of dollars, and inestimable trips down intellectual rabbit holes that would end up being dead ends.

Applying AI

Unilever, for instance, has harnessed AI to launch new products, including Hellmann’s Plant-Based Mayo. Ocean Spray is using AI to plumb all the possible healthful effects of cranberries, including ones that may be influenced by various locations where they’re grown.

Coca-Cola used AI late last year to come up with a new limited-edition product it called Y3000 Zero Sugar, a flavor that had a tutti-frutti character to it. To help market the concoction “from the future,” the beverage giant used an AI program to create a new version of its iconic Christmas polar bear mascot.

And PepsiCo product developers noticed that people were interested in immunity, thanks to an AI tool that analyzed millions of social posts, recipes, and menus. That led R&D to develop water options with immunity-boosting ingredients under its Propel label.

The vast majority of new food products still fail, so the industry increasingly is grasping AI techniques not only to help R&D efforts to fail faster but also to find promising ideas much more quickly, notions that often come complete with road maps showing how to make them into products.

“Our combination of data, computational power, AI, and automation is driving advances at astonishing speed,” says Manfred Aben, a global vice president in science and technology for Unilever. “We’re able to discover what humanity never had the tools to discover before, advancing scientific fields not by years, but by decades.”

The most important applications of AI to food so far are in R&D and marketing, where the raw power of digital computation and algorithmic connection is yielding game-changing capabilities for CPG companies to come up with innumerable new ideas for products and how to sell them.

I used AI in combination with what’s trending, what I’ve seen myself in the world, and how to use food science with what’s trending.

Doing More

“It’s all about capacity—the capacity to do more,” says Vinay Indraganti, founder and CEO of AI developer BCD iLabs. “New product development has dropped since COVID, but demand has gone through the roof for changes and to meet trends in products. Yet [product developers] spend a lot of time doing nonessential and non-value-added work, and AI can help that.”

In fact, says Abhinav Agrawal, partner and managing director with consultancy AlixPartners, “in terms of getting value that can be quantified at this point, manufacturing and supply chain applications are leading the pack, while from an interest and trial point of view, product R&D is taking the lead” in the food industry.

Food product developers’ early embrace of AI is changing their industry’s reputation for “being followers” when it comes to new technology, says Tim Gaus, smart manufacturing business leader for Deloitte. “Specifically, a lot of the leading in AI is going on within the food space. And that’s a strong sign of the health of that sector.”

The digital world is tailor-made for AI to boost the food business. Food yields more publicly available information than just about any other category on the internet and is one of the most-discussed topics in social media, blogs, recipes, and other forums, providing a data-rich environment for AI algorithms to explore and intertwine information that it fashions into human-like insights. Even the COVID-era boom in home-delivery food services, with their detailed menus, provided a serendipitous boon to data for AI.

Still, Agrawal cautions, AI “is not a panacea. It’s good at assisting researchers, but in the current format, it hasn’t eliminated the need for R&D. And while it can offer nuances, it’s not all the way there.”

A lot of the leading in AI is going on within the food space. And that’s a strong sign of the health of that sector.

AI is rapidly evolving into a world-changing technology, having written two main chapters so far. The first was machine learning and “foundation” AI models, which food companies have used for several years to plumb internal data for insights.

The second era began last year when OpenAI launched its ChatGTP generative AI program, which accesses billions of data points in publicly available external information to come up with raw observations, analysis, insights, and recommendations in the blink of an eye. The results can range from ingredient predictions to details necessary for personalized nutrition to trend analyses.

“Infused with gen AI, you can have 100-plus ideas but also fully written concept statements fueled by consumer information, and can render them not only with imagery on the packaging and on the plate but also with the consumer context, and a presentation—whether it’s casual entertaining or a Super Bowl party,” says Justin Shimek, CEO and chief technical officer of Mattson, an insights and product development firm. Mattson has developed its own proprietary AI tools to help support product development initiatives.

The most powerful applications combine different types of AI. “Manufacturers can take that second- and third-party data and assimilate it with their [proprietary] data into different models and see trends, and be able to do formulations of things that used to be unheard-of,” says Jeff Van Pelt, principal, consumer industries, at industrial technology provider NTT DATA Business Solutions. “And can we dig back into our archives to come up with a formulation that wasn’t relevant two years ago but now is extremely relevant? AI enables them to do that in ways that were unimaginable before.”

Kraft Heinz Tackles Plant-Based

In trying to deliver successful plant-based foods amid the minefield of failed efforts in that arena, Kraft Heinz agreed two years ago to form a joint venture with NotCo for the Chile-based company to leverage its AI engine, called Giuseppe, to come up with superior products that would exploit the CPG giant’s iconic animal protein–based brands.

“We wanted to find new ways to deliver plant-based foods that don’t ask consumers to drastically shift their behaviors,” says Robert Scott, president of research and development for North America for Kraft Heinz. The joint venture declared that it would “leverage the inherent strengths of both companies” to do so.

“This AI technology gives us the ability to understand plant-based foods at the molecular level so that our recipes use unexpected combinations of plant-based ingredients to mimic the flavor, color, smell, and functionality of animal products,” Scott says.

Giuseppe gauges if there are categories where plant-based products could play and then determines whether Kraft Heinz has the brands to fill that spot. Delivering a product with the taste that consumers would demand is paramount, especially among “flexitarians” who eat both plant- and animal-based items. The price points must be low enough to attract mainstream shoppers.

With the promise of more products to come soon, so far The Kraft Heinz Not Company has introduced Giuseppe-inspired KRAFT NotMac&Cheese, KRAFT NotCheese Slices, NotMayo, Oscar MayerNotHotDogs, and Oscar MayerNotSausages. Scott reports fast sales takeoffs for these products, even in categories such as plant-based meats, where American consumers have become dubious.

AI technology gives us the ability to understand plant-based foods at the molecular level so that our recipes use unexpected combinations of plant-based ingredients to mimic the flavor, color, smell and functionality of animal products.

Unilever Speeds Up Iteration

Unilever has more than 500 AI projects globally as it uses what the company calls “in silico testing,” which are experiments performed via computer simulation, particularly to test millions of recipe combinations in seconds. That “helps us design products to suit regional taste preferences,” Aben says. “We run virtual tests and scenarios to optimize products before the lab and scale-up stage, bringing efficiency and cutting time to market.”

After that, Unilever R&D hubs in Shanghai and Mumbai use real-time consumer data to develop new insights, followed by rapid prototype development and testing of those models via digital commerce in a matter of days. “We’re unleashing new levels of efficiency in our operations, which is in turn freeing up our R&D experts to deliver impactful innovations,” Aben says.

For instance, Unilever scientists developed Knorr Zero Salt Stock Cubes after completely reinventing the typical bouillon cube by testing and digitally screening thousands of options with AI to craft a blend of vegetables and herbs, and to create a rich flavor with the texture and structure of a traditional stock cube—but with zero salt.

And to develop Hellman’s Plant-Based Mayo, Unilever harnessed AI models to replace the egg emulsifier with a plant-based alternative without the need for multiple recipe testing and traditional product development trials.

Brightseed Unearths Bioactives

Brightseed was an AI pioneer with a proprietary computational platform called Forager that is building what the company describes as “the world’s largest digitized library of natural compounds mapped to human health targets based on public literature and original datasets.”

“The AI makes connections you wouldn’t do alone as a researcher,” says David Brown, vice president of business development for Brightseed. “It’s hard enough to research a single molecule versus seven million. This can create massive shortcuts in product development.”

Brightseed collects thousands of plants from all over the food chain, and all over the world, and puts them through natural product chemistry and metabolomics, including mass spectrometry. Forager creates a digital, molecular blueprint of each one of those plant samples in order to identify bioactive compounds and predict their unknown impacts on human health.

“It de-risks and saves a lot of time,” Brown says. “Instead of hunting for a needle in a haystack, it’s like looking for a bowling ball in a haystack.”

The company’s first product on the market contains two bioactive compounds that are central metabolic regulators to help with gut strength and gut linings—and that were plucked by Forager out of the waste stream from hemp seed. Brightseed now is conducting multiple clinical trials and plans to bring more products to market early next year.

Instead of hunting for a needle in a haystack, it’s like looking for a bowling ball in a haystack.

Meanwhile, clients including Ocean Spray, ADM, and Blue Diamond “use us to do profiling of their ingredients to find out what new compound we can find that then will help target health properties,” Brown says. The goal, he says, is “finding the good stuff they didn’t know was there.”

Ofi (Olam Food Ingredients) also has turned to Brightseed to flush out ideas for features and products, and two years ago jumped into a partnership to identify bioactives in black pepper and garlic. Speed, the company believes, is of the essence.

“A lot of times, people wait for everything to fall into place,” says Kamesh Ellajosyula, ofi’s president and chief innovation and quality officer. “It’s important [with AI] to engage early. It’s the biggest opportunity for a lot of us in the food business.” ft