This is the conclusion of our three-part series on AI in the food industry! Check out Part I: Food Formulation and Part II: Operations, along with our latest Future of Food Series: Food x AI video that ties the three parts together.
Bluestein’s Take
Consumers are more demanding and discerning than ever about their food choices due to changes in demographics, the ubiquity of information, social media, as well as global and cultural movements.
AI has potential to help companies keep up and empower consumers to intelligently manage their diet and lifestyle, harnessing messy and noisy consumer data to unlock key insights.
We are most excited about solutions that integrate several data sources seamlessly, generating insights and personalized recommendations to delight customers and solve key problems – and about the promise of 1:1 marketing at scale to help companies better engage their customers.
The Fickle Foodie Phenomenon
Modern consumers are rapidly changing their demands – becoming increasingly discerning about their food choices – and companies need to keep up. These demands include higher nutrient density, less processed ingredients, more functional benefits – and convenience above all. For example, consumers report checking ingredient and nutrition labels 43% more than three years ago,1 and a survey by good natured showed that 72% of respondents were “regularly buying or increasing their consumption of ready-to-eat meals, meal kits, and takeout food and delivery.”2
What’s driving these changes? Demographics, the ubiquity of information, social media, as well as global and cultural movements. For one, the US population is becoming increasingly racially diverse, with 51% of the under 18 population identifying as a race other than solely white.3 Younger consumers in the 18-to-24 cohort are also more likely to avoid certain ingredients (most frequently red meat, dairy, and gluten), and report higher rates of food allergies and intolerances.4 In addition, the US population has become older, with 16.8% over 65 years old in 2020 versus 12.6% in 1990; we’ll likely continue to see changes in consumer behavior from this cohort as they look to increase their health span.5 These trends indicate a shift towards more diverse food choices, reflecting the changing composition of the population.
Social media, along with the general ubiquity and proliferation of information, has also significantly influenced consumer expectations. Modern consumers are simultaneously more confused yet more invested in food than ever. Conflicting opinions on food choices have made consumers less able to parse signal from noise. They’re also more fickle, eager to try the latest food trends that go viral on TikTok (see freeze-dried candy and Dalgona coffee) or to incorporate new products endorsed by their favorite influencers. Brand loyalty has therefore declined; 57% of Gen Z Americans are less loyal to brands now than they were pre-pandemic, and more than 1 in 3 US customers are not loyal to brands.6
These trends have created a need for companies to respond swiftly with both products and content that resonates to capture and retain limited consumer attention. Companies spend on average two years developing a new food product and 15,000 are launched annually, yet there’s a staggering 70-90% failure rate.7 In a world where customer acquisition costs are high and brand loyalty is low, brands need to focus on providing the best products and experiences to attract and retain customers – especially as it relates to the younger demographics.
AI has the potential to help companies gain and maintain a competitive advantage with the diverse consumer. This includes the ability to better target, engage with, and personalize offerings for consumers; identify, analyze, and capitalize on trends; and optimize product development to lower the failure rate for new products.
Dishing Out Innovation: AI's Menu of Opportunities
AI has vast potential to help businesses adapt and respond to consumer needs, while simultaneously empowering individuals to make more informed, tailored decisions about their diet and lifestyle. From enabling deeper consumer insights to offering personalized nutrition plans for individuals, AI is reshaping the future of food consumption and production.
The Digital Palate: AI's Taste for Consumer Satisfaction
AI can help companies meet consumer demands by enabling deeper insights, rapid feedback, more agile product development, and the ability to build stronger engagement and loyalty via more personalized content. AI is capable of analyzing and interpreting a rich tapestry of data from variety of information sources, including purchase patterns, dietary preferences, cultural influences, and even psychological factors, that can help companies in the food industry gain unprecedented insights into consumer behavior.
AI also enables rapid product feedback for companies in the food space, reducing the cost and time. AI-powered tools include sentiment analysis of social media posts, real-time data from smart kitchen appliances, or even AI-driven taste prediction models – which can be coupled with biofeedback from wearables or even brain imaging to better reveal excitement for products, crave-ability, and nutraceutical or health effects.
An accelerated feedback process allows for more agile product development, quicker adjustments to market trends, and ultimately, a more responsive and consumer-centric approach to food production and marketing. Companies can stay ahead of consumer demands, fine-tune their offerings more efficiently, and maintain a competitive edge in the ever-evolving food landscape. One example of this is AB InBev’s Beck’s beer brand that launched Beck’s Autonomous, the world’s first beer and full marketing campaign made with artificial intelligence, which was selected by AI as the “favorite among millions of different flavor combinations it generated.”8
We’re living in an attention economy. For companies trying to create memorable and differentiated brands, AI can be a powerful tool to foster deeper engagement and loyalty, as well as offering the ability to tailor content to consumers such as custom product recommendations or targeted messaging. AI can help companies understand brand perception in real-time, allowing for swift adjustments to marketing strategies, as well as create high quality, personalized content at lower costs. For example, Starbucks has perfected personalized promotions using AI on their mobile app to elevate loyalty and retention.9 Companies in the food space can therefore craft more cohesive, responsive, and resonant brand strategies, reducing the risk of brand dilution and ensuring they maintain a strong market position amidst changing tastes and preferences.
Rise of the AI-Powered Consumer
AI can help consumers directly by making personalized nutrition more accessible, tailored, and effective, catering to the nuances of individual dietary needs, cultural preferences, allergies, aversions, and health goals. AI-powered apps and devices that analyze eating habits, activity levels, and even genetic predispositions can create highly personalized meal plans and nutritional recommendations with higher accuracy and without as heavy a manual lift from consumers. Computer vision has the power to identify contents and calorie counts in food automatically, and wearables like glucose monitors can provide further insight into health data that previously wasn’t tracked. In Japan, Nestlé piloted a “Wellness Ambassador” program, an AI-powered personalization platform that analyzed users' DNA and blood samples, along with pictures of their food, to provide personalized nutrition recommendations and suggest supplements.10
AI can help consumers navigate meal preparation, grocery shopping, and food delivery, a tedious task families face. AI can help to suggest healthy recipes from a picture of a household’s fridge or a few random ingredients, online platforms can use AI to personalize recommendations and promotions to simplify and customize ordering, and easy collection and centralization of data can create an integrated system for health management through tailored suggestions that can be automatically selected and delivered to the home.
In addition, AI can help analyze, consolidate, and present information about a product beyond the nutrition label to help consumers make a more informed choice. AI can present the health and environmental impact of ingredients – and make recommendations accordingly that are specific to a consumer’s dietary preferences, restrictions, allergies, and health goals.
AI's Blind Spots: When Algorithms Miss the Mark
AI solutions in the consumer sector face challenges ranging from data quality issues to model reliability concerns. Companies face the complexities of translating digital trends into real-world consumer behavior and must hold their models to a high standard of accuracy and reliability to be useful.
Low quality and incomplete data inputs
Social media is useful for food companies looking to understand consumer preferences. Yet the sheer volume of posts about food is overwhelming and a lot of it is low quality. Social media can amplify fleeting trends that are more about novelty and shareability than actual sustained consumer interest driving purchase behavior. For example, the "Butter Board" trend went viral on TikTok in 2022, garnering millions of views and inspiring countless videos. However, it didn't lead to significant long-term changes in consumer buying habits or restaurant menus, highlighting the gap between viral social media content and actual consumer behavior.11[11]
Incomplete and fake data is also a risk. Only a small subset share their food experiences online, and those who do might not be representative of the broader consumer base. Relying too heavily on data from specific apps or devices can create blind spots. Fake accounts and reviews, bots, and paid influencers can also skew the perception of what's genuinely popular, potentially leading businesses to invest in product development or marketing campaigns based on trends that aren’t as prevalent in the real world. For instance, in 2022, Yelp removed more than 700,000 reviews and a 2013 study found that 20% of Yelp reviews were fake (certainly there are more now over ten years later), demonstrating the scale of fake data that can influence consumer perceptions and business decisions in the food industry.12 A 2018 study by Ghost Data estimated that up to 95 million Instagram accounts could be bots, which can significantly skew perception of food trends and influencer marketing effectiveness.13[13]
Model reliability and usability
One of the most significant concerns is the tendency of AI to produce "hallucinated" results – outputs that seem plausible but are inaccurate or entirely fabricated. Hallucination can be hard to spot because models do not say or highlight if they are ‘unsure’ about something being true. For example, Google’s Gemini LLM model came under fire for recommending users add glue to their pizza to prevent cheese from sliding off that was generated from a Reddit comment meant as a joke.14 LLMs can do a decent job generating a baseline recipe, but struggle with any edits to the recipe, often failing to adjust proportions of other ingredients and cook time, like OpenAI's GPT-3 modifying a chocolate chip cookie recipe to be gluten-free by replacing wheat flour with almond flour without adjusting other ingredients, potentially resulting in a botched product.
When a company is using consumer data to make marketing, sales, and product development decisions worth thousands or millions of dollars, accuracy is paramount. Companies can mitigate these risks by thoughtfully training Small Language Models (SLMs) and implementing robust guardrails including human oversight, fact-checking mechanisms, or AI systems trained specifically to detect and filter out inconsistencies. SLM are less likely to hallucinate, given they are trained on specific, relevant datasets, and use less compute power to operate – but they aren’t as accurate or powerful, facing drawbacks that include potential for bias, data privacy concerns, limited knowledge base, specialized performance, and limitations in handling complex datasets.
Even when an AI model is functioning optimally, the problem it is solving must be useful to its user. For example, an AI model may generate an ideal meal plan for a person, but if that person’s primary issue is compliance and accountability, the said meal plan is basically useless. AI systems need to be designed with a first principles, customer centric approach in mind to create value.
Mapping the Consumer AI Frontier
Market Segments
To better understand AI in the consumer landscape, we created a market map of key players in the industry. The chart below maps out companies across several pillars of the industry, from consumer insights to consumer purchasing.
Consumer Insights: Companies in this segment, specialize in understanding consumer behavior and preferences to generate insights across different departments like marketing or R&D. Data from social media tends to be more abstract (e.g., identifying a trending flavor) but is easier to collect, while data from targeted consumer feedback is more specific (e.g., sensory preferences of a specific demographic for a particular product) but requires more effort to gather. For example, Tastewise uses AI to analyze social media and restaurant data, providing real-time insights into consumer trends and preferences, while Thimus uses neuroscience techniques, like EEG data on brain activity, along with traditional sensory analysis, to generate better data and insights from consumer feedback.
Product Information: AI companies can help develop personalized meal plans based on accurate nutritional information. For example, Verdify helps customers personalize recipes and meal plans to their own tastes, and then works with brands and retailers to provide customer insights, and Foodgraph is creating the largest catalog of US food products to provide a single source of truth for brands and retailers.
Marketing: Companies use AI to enhance marketing strategies and consumer engagement. These platforms help brands target their audience more effectively and personalize their marketing efforts. For example, Spoon Guru uses AI to personalize food recommendations based on individual dietary preferences and restrictions. By leveraging AI, these companies enable brands to deliver more relevant and impactful marketing messages to their consumers.
Consumer Purchasing
The consumer purchasing segment encompasses AI-powered solutions that help consumers in making informed decisions about products and services, especially related to food, health, and wellness. The segment is divided into three subcategories: Search and Discovery, Meal Planning, and Personalized Health. The market demonstrates increasing customization in the consumer sector, driven by the growing demand for health, wellness, and convenience. By harnessing AI, these companies are not only meeting current consumer demands but also paving the way for future innovations that will further enhance personalization and improve overall well-being.
Search and Discovery: Companies focused on helping consumers discover new products and make informed purchasing decisions, providing personalized recommendations and product information to enhance the shopping experience. For example, Yuka analyzes product labels and provides a health score, helping consumers choose healthier options while Fig can help customers navigate complex diets like low FODMAP or celiac friendly.
Meal Planning: AI companies offer solutions to simplify meal planning and preparation. Consumers can save time and make healthier choices through individualized meal plans, recipes, and grocery lists. For example, Relish helps people discover recipes, weekly menus, shopping lists, and order groceries all in one place.
Personalized Health: Companies use AI to analyze individual health data and provide personalized recommendations for diet, exercise, and lifestyle. For example, Viome uses microbiome analysis to offer tailored dietary advice, while InsideTracker analyzes blood, DNA, and wearables data to provide health insights.
Evaluative Criteria
AI has significant potential to personalize the consumer sector. We believe that the best solutions in the space are those that provide:
Flywheel effect around personalization: The best AI solutions in the consumer space should offer highly personalized experiences. This includes tailoring recommendations, content, and interactions based on individual user preferences, dietary needs, health goals, and past behaviors. The ability to adapt and learn from user interactions over time is crucial for maintaining relevance and engagement. The solution must get quickly and demonstrably better with more usage.
Multimodal, user-friendly interface: AI solutions should be seamlessly integrated into intuitive, easy-to-use interfaces. The technology should enhance the user experience without adding complexity. Interfaces should be multimodal, giving consumers flexibility to input data via talk, text, images and video, and receive similarly diverse outputs where applicable.
Data privacy and security: With the sensitive nature of personal health and dietary information, top-tier AI solutions must prioritize robust data protection measures. This includes secure data storage, transparent data usage policies, and compliance with relevant regulations, especially the recent European ‘AI Act’. As AI solutions become more intelligent, companies need to be prepared to handle insights and correlations around sensitive categories like pregnancy, sexual orientation, and illness (for example, in 2012, Target predicted that a high school girl was pregnant based on her purchase information before even her father knew).15
What’s Next
We’re excited about AI in the consumer space, both from the perspective of investors and as consumers ourselves. We are most excited by solutions that can integrate many types and sources of data and then optimize nutrition and planning to help people reach their health goals, and companies enabling 1:1 marketing at scale by intelligently segmenting customers and generating customized content. If you’re an entrepreneur in this space or know of one, we’d love to hear from you.
That’s the conclusion of our three-part series on AI in the food industry! Check out Part I: Food Formulation and Part II: Operations, along with our latest Future of Food Series: Food x AI video that ties the three parts together.
Thank you to our Analyst Aaditi Tamhankar and our PhD fellow Tarini Naravane for leading the efforts on this series.