KANSAS CITY — Artificial intelligence (AI) is emerging as a valuable tool for food and beverage makers looking to bolster front-end innovation. Manufacturers, restaurants, ingredient suppliers, flavor houses and more are leveraging insights from machine learning to get closer to consumer trends and market more nuanced propositions.
New food and flavor concepts traditionally have been ascribed to culinary experts, chefs and product developers, said Ron Harnik, vice president of marketing at Tastewise, an Israel-based AI food and beverage platform. Translating an idea into a finished product can take months or even years.
“The processes that are set up to take products to market simply aren’t built to be quick and accurate enough to reflect how fast consumers are changing,” Mr. Harnik said. “Companies rely on a lot of outdated sources, like surveys, focus groups and retail data.”
Tastewise pulls data from social media, home cooking analytics and millions of restaurants and their menus to equip users with a deeper understanding of real-world eating and drinking moments. Casting such a wide net, and making sense of the findings, wouldn’t be possible with traditional research methods.
“If you do a focus group with 100 people, all you’re going to learn is that 100 people want 100 different things,” Mr. Harnik said. “You have to start looking at millions of people to see patterns.”
Research papers, consumer conversations, grocery platforms and expert food blogs are just some of the sources Minneapolis-based Spoonshot uses to deliver personalized insights to food and beverage makers. Those mountains of information aren’t useful on their own, though. The real value in AI is its ability to transform massive amounts of data into coherent, relevant and actionable insights.
“We’re in a world where data is growing exponentially,” said Kishan Vasani, co-founder and chief executive officer of Spoonshot. “Something like 90% of the world’s data was created in the last two years. How do you navigate that and get to the truth? You need someone to help take the friction away.”
Spoonshot feeds data from more than 28,000 sources into its Food Brain database. A team of culinary experts work with engineers to codify their knowledge and input it into the algorithm, enabling Food Brain to draw connections between seemingly disparate data points using the domain of food science.
Codifying that expertise to unify data can spark answers to questions users never even thought to ask, or as Mr. Vasani called it, “unintuitive intelligence.”
As an example, he imagined several events occurring on the same day: A celebrity chef posts a new dish to Twitter, a startup launches its first product and an academic journal publishes research on novel food processing technologies.
“Each of those could be understood and reported back to users on their own,” Mr. Vasani said. “What’s powerful about Food Brain is that it finds the signals in all that noise. It’s doing computation on a high scale, evaluating the context of every user and giving you insights based on everything that’s happening.”
PepsiCo, Unilever, Mars, Danone, Campbell Soup, TreeHouse Foods and Kraft Heinz are among the growing list of CPG companies using AI to shape their innovation pipelines. Suppliers like Cargill and Givaudan also are using AI to identify ingredients and flavors they can sell to manufacturers.
Freshly, a meal kit company owned by Nestle, used Tastewise’s platform to research consumer relationships to comfort food and global cuisine. AI-generated insights sparked the launch of a golden chicken with apricots dish, one of the first Freshly meals to introduce international flavors.
“The success of the dish was in large part due to Tastewise insights and being able to use the tool to find a balance between newer flavors/ingredients and sides, flavors and ingredients that we could pair with them that would make the dish more approachable,” said Rachel Waynberg, a meal innovation leader at Freshly. “What used to take three days of painstaking research took three hours of data-driven analysis.”
Trend predictions and consumer insights are two common uses for Spoonshot’s platform. Companies also use Food Brain for ad-hoc research, identifying strategic innovation opportunities as well as specific product concept recommendations.
“Some people are on an open-ended journey of discovery,” Mr. Vasani said. “Others come with more tactical questions, like ‘What flavor should we do next?’ or ‘Who is our top competitor for this SKU, which one is winning and why?’ Not a month goes by that I don’t hear about a new use case.”
Applications for AI in front-end innovation are still evolving, and the technology shapes only a small portion of products today. Mr. Vasani estimated just 5% of new food and beverage products incorporate AI insights at the decision stage.
"There are approximately 20,000 product launches every year in America," he said. “If you buy in to the story that anywhere from 50% to 85% of new product launches fail within six months, then the amount of resource waste and inefficiency is a big problem for the industry."
AI is helping companies identify whitespace and bring on-trend products to market faster. It’s also offering a higher degree of confidence about how consumers will respond to a new product and how it will perform among specific groups.
New York-based Analytical Flavor Systems (AFS) uses AI to model perceptions of flavor and texture. The company quantifies sensory adjectives and consumer language, drawing on a diverse collection of products and consumers from around the world to train its Gastrograph tool. Just 10 to 15 tasters need to review a product for Gastrograph to predict how any other consumer demographic will respond to it.
AFS last year partnered with Ajinomoto Co. for a blind study comparing its predictions with those generated through conventional research methods. Gastrograph used a dozen tasters in Japan to predict consumer perceptions and preferences across different demographic groups in China. Meanwhile, an independent research firm in China conducted central location testing (CLT), surveying hundreds of consumers about the same product.
“Although CPG brands have relied on time-consuming CLT data, our first publicly available validation study showed what we already knew: AI can predict consumer tastes far faster, and even more accurately,” said Jason Cohen, founder and CEO of AFS.
The AI platform exceeded researchers’ expectations by delivering accurate predictions in less than two weeks, added Hiroya Kawasaki, associate general manager at Ajinomoto’s Institute of Food Sciences and Technologies. The CLT test took two months to complete, making Gastrograph “at least an order of magnitude faster than existing empirical methods,” he said.