Artificial intelligence is moving from a supporting role in food innovation to a more embedded position across R&D, formulation, supply chain planning, sensory prediction, and consumer insight. For ingredient suppliers, the opportunity is not simply to automate tasks, but to connect data, science, and application know-how in ways that compress development timelines, improve predictive accuracy, and help manufacturers respond to fast-moving market pressures.
For IFF (International Flavors & Fragrances) and Corbion, AI’s value lies in its ability to strengthen — rather than replace — the expertise of scientists, flavorists, and application teams. The technology is helping companies interpret complex datasets, model multiple formulation variables simultaneously, and translate emerging consumer signals into more targeted product development.
From trial-and-error to multi-variable modeling
At Corbion, AI is becoming “an integral part” of how the company approaches ingredient performance across applications, says Symone Kok, project management director.
“In R&D, we are using AI to analyze large datasets that capture how our fermentation-based ingredients behave across different formulations, processing conditions, and food matrices,” she explains. “This allows us to identify performance patterns and optimization opportunities that would take significantly longer to surface through conventional trial-and-error experimentation.”
The ability to analyze large datasets more effectively is particularly relevant in complex food matrices, where ingredient behavior is shaped by multiple interacting variables.
“The real value lies in the ability to work across variables simultaneously,” Kok says. “Where traditional methods require sequential testing, AI enables us to model interactions across multiple parameters at once, accelerating our understanding of how ingredients can be fine-tuned to deliver the best functional outcomes in specific applications.”
“This does not replace deep scientific expertise — it amplifies it.”
AI helps ingredient teams move beyond trial-and-error by modeling multiple formulation variables at once.
Connecting foresight, science and formulation
IFF is taking a similarly integrated view. Benjamin Mattei, VP intelligence & digital transformation, Taste, says the company’s R&D strategy is focused on “connecting foresight, science, and formulation from the earliest stages of research.” This is to help flavorists and application designers deliver innovations that are relevant and technically robust.
“AI is embedded throughout the process — insight development, concept creation, testing, validation, and supply, so it’s an end-to-end ecosystem,” he says. IFF teams use digital flavor solutions, predictive modeling, and multi-sensory design tools to explore ideas, optimize flavor performance, and tailor profiles to customer goals.
Among these tools is Concept Muse, “an AI-powered engine built on IFF’s proprietary insights and taste technology, designed to deliver faster innovation driven by real-time signals.”
For IFF, one of AI’s practical advantages is improving “first time right” innovation. Mattei says the technology enables quicker decision-making and helps customers respond more rapidly to market shifts. This is a priority as brands navigate volatile costs, evolving dietary needs, and increasingly fragmented consumer expectations.
Turning weak signals into product outcomes
Consumer insight is another area where AI is changing the innovation process. Rather than treating trend analysis as a separate front-end exercise, companies are linking preference data more directly to formulation decisions.
Mattei notes that “there are thousands of signals shaping today’s food and beverage landscape and giving us clues about behaviors, priorities, and ingredient shifts.”
IFF’s intelligence function analyzes online consumer trends, market activity, and ingredient performance to identify where demand is forming and why. “AI plays an invaluable role in this process, enabling us to detect patterns and identify weak signals early, turning them into clear, actionable insights,” he says. “But we’re moving beyond simply analyzing consumer data to actively embedding those insights into how we formulate and innovate.”
That means large-scale preference data can be brought into intelligent formulation workflows, helping teams translate emerging signals — from evolving flavor expectations to GLP-1-related health priorities and dietary needs — into product outcomes. Formulation pathways can then be assessed against cost, regulatory, and allergen constraints with greater precision.
Symone Kok: Corbion is using predictive modeling to support Listeria risk management and reduce food waste.
Corbion is also combining market-facing and customer-facing data streams. “On their own, consumer datasets tell you what people say they want. Customer data tells you what manufacturers are actually grappling with in production,” Kok says. “When you bring those two streams together, you start to identify convergence points, areas where a consumer-facing trend is likely to create a formulation or food safety challenge downstream.”
This gives ingredient suppliers a more proactive role. Instead of waiting for a formulation brief, Corbion can bring early-stage technical insight to customers and anticipate where production, safety, or shelf life issues may arise. “AI makes that kind of anticipatory engagement more scalable,” Kok adds.
Sustainability through prediction and prevention
Sustainability is another major proving ground. AI’s ability to model outcomes before physical trials can reduce wasted time, materials, and energy in R&D. But for ingredient suppliers, the sustainability case also extends into safety, shelf life, sourcing, and carbon impact.
At Corbion, Kok points to two standout areas. The first is food safety and pathogen control, specifically Listeria monocytogenes risk management. “AI-powered predictive modeling allows us to simulate microbial outgrowth scenarios under real production conditions, giving manufacturers the ability to assess risk and adjust formulations before issues arise in the field,” she says.
“This has meaningful implications for food waste reduction: avoiding a recall is not only a safety outcome, it is also a sustainability one.”
The second is validation speed. Corbion’s Listeria Control Model integrates AI-driven approaches to shorten the time required to validate preservation solutions. According to Kok, this helps move from lengthy challenge studies to model-supported validation that is “scientifically robust and faster to market.”
In a sector where safety, speed, and waste reduction are closely linked, such tools could become increasingly important to manufacturers looking to de-risk innovation.
IFF is applying AI to sustainability through formulation and data integration. Mattei says the company is centralizing sustainability data from suppliers, industry benchmarks, public databases, and internal metrics to build AI-powered engines that can predict carbon footprint impact during flavor formulation. The objective is to preserve flavor performance and differentiation while ensuring solutions meet customer standards and IFF’s own responsible growth expectations.
“We’re also working to integrate biotechnology with chemistry, data science, and application engineering to develop high-performance ingredients and materials that support efficient, scalable, and sustainable manufacturing,” Mattei says. “This helps us contribute to the growing bioeconomy, as well as enabling solutions that deliver value for customers, consumers, and society.”
Governance, transparency and human oversight
As AI becomes more deeply embedded, transparency and governance are rising up the agenda. Food innovation is highly regulated, and AI-generated outputs must be explainable enough to support safe, compliant decision-making.
Corbion has established internal AI governance policies and ensures qualified representatives across the business stay current on evolving regulatory and safety guidance. “This is not a static exercise,” Kok says. “The regulatory landscape around AI in food and life sciences is moving quickly, and maintaining that awareness requires structured, ongoing effort.”
AI is shifting from isolated pilots to core innovation infrastructure across food ingredients.
Internal AI literacy is equally important. “We invest in training our teams to understand both the capabilities and the limitations of AI tools, so that the technology informs decisions rather than substituting for scientific judgment,” Kok says. “The culture we are building is one of collaboration between human expertise and AI capability, where our scientists remain in the lead, with AI as a powerful instrument in their hands.”
IFF’s digital transformation is also grounded in human oversight. Mattei stresses that AI can support sourcing, formulation changes, production planning, and disruption forecasting, but “accountability and decision-making remain firmly with our experts.”
He identifies three foundational values: people, ethics, and sustainability. “For us, AI is not to replace people but to empower them — it is an enabler. Like a catalyst, AI can accelerate outcomes, but it is human creativity and expertise that define the objective, shape the approach, and ultimately determine success.”
On ethics, IFF emphasizes rigorous data standards, GDPR (General Data Protection Regulation) compliance, and oversight through its AI Trust Council. On sustainability, Mattei says the company deploys AI “only when it delivers clear, differentiated value,” ensuring it complements efficient or more sustainable alternatives rather than becoming technology for technology’s sake.
From isolated pilots to innovation infrastructure
Recent Food Ingredients First coverage has shown how quickly the conversation is evolving. Cargill has framed AI as a capability embedded across the full value chain, from on-farm tools and supply chain optimization to customer co-creation and formulation. The company’s approach underlined a broader industry shift: AI is moving beyond isolated pilots and becoming part of core innovation infrastructure.
That shift is also visible at the consumer-facing end of the market. Unilever Foods told us how brands, such as Knorr and Hellmann’s, are adapting to AI-powered search, voice tools, and large language models. They are also applying digital testing and simulation to accelerate innovation cycles. Its work on Knorr’s Fast & Flavourful Paste helped halve development time by digitally exploring formulations before physical trials.
Benjamin Mattei: IFF is embedding consumer signals into formulation workflows to improve “first time right” innovation.
Meanwhile, Food System Innovations’ Food Intelligence Lab is building open-source AI tools and datasets for sustainable protein development. The initiative combines sensory data, instrumental testing, and AI models to shorten R&D timelines and help developers move closer to animal-based benchmarks for taste, texture, and functionality.
Together, these examples point to the same conclusion: AI’s greatest value in food innovation is not autonomy, but augmentation. It helps teams screen more options, interpret more signals, and reduce trial-and-error, while keeping human expertise central to decisions around safety, sensory quality, scalability, and commercial relevance.
Partnerships without losing customer intimacy
The same balance applies to partnerships. IFF and Corbion both see external collaboration as important, but not as a substitute for deep application expertise or customer intimacy.
“We actively explore partnerships across start-ups, academia, and technology platforms to complement our internal capabilities and stay at the forefront of emerging developments,” Mattei at IFF says. “The key is balance, leveraging external innovation where it adds speed and scale, while anchoring everything in our deep in-house expertise.”
Kok echoes this pragmatic approach. Corbion prioritizes collaborations that address real customer challenges — improving shelf life, reducing food safety risks, accelerating time to market, or optimizing formulation performance under specific processing conditions.
“In our industry, the most meaningful breakthroughs come from a deep understanding of each customer’s unique context, their processes, constraints, and end markets,” she says. “Our approach, therefore, is to use AI and external partnerships to enhance our ability to listen, anticipate, and co-develop solutions with customers, not to replace the human connection that underpins long-term value creation.”
