Tackling global food challenges with AI: insights from an R&D leader

How AI can assist with decision-making

“One key aspect that sets us apart from the rest of the animal kingdom is that we can differentiate between right and wrong,” Abhishek says. “At the same time, we make thousands of decisions daily, both in our private and professional lives. I see my job as making it easier for people to make those decisions, using AI as a tool to help lighten that load by offering clarity quickly.” On one level, it’s about reducing the number of decisions we need to make by delegating them to a copilot “or an AI version of you that can help automate, augment or assist in this decision-making, or minimize the uncertainty around whether you made the right choice or not,” he says. “We are always shooting in the dark. We constantly wonder if we made the right decision. Our job is to alleviate some of those concerns.”

The accidental data scientist

Abhishek’s journey into AI was not linear. Today, in addition to leading R&D AI at Cargill, he teaches AI and data engineering at the University of St Thomas opens in new tab/window in Minnesota. But when he graduated with a degree in electronics and telecommunications, he didn’t have a clear direction. “After college, like many other young, wide-eyed people, I also wanted to explore the world and see what else I could do,” he recalls.

A fortunate first job in business consulting introduced him to the discipline that would become his passion — a field at the intersection of behavioral psychology, mathematics and statistics. “At that time, it was called business intelligence — BI. Now we call it data science and AI.”

Curious about human behavior and decision-making, Abhishek pursued opportunities across various industries: retail, CPG (consumer packaged goods), pharma, finance and more. “I never liked being boxed in; I’ve always wanted to connect the dots across disciplines,” he says.

“I’m fascinated by understanding how people operate: ‘This is how we do things in molecular biology.’ Yes, I love data, but I also have a genuine curiosity to explore uncharted areas and learn about the people and processes within them. It’s that combination that I truly love — taking action to solve the problem at hand.”

Now, as the leader of AI for research and development and innovation at one of the world’s largest food companies, he views Cargill as a place brimming with challenges — and possibilities. “It’s like being a kid in a candy store,” he says. “There’s always something new to learn and something meaningful to improve.”

Small changes, big impact

Across Cargill, AI is already delivering results. In R&D, generative AI accelerates product innovation by learning from past formulations and experimental data. These aren’t futuristic concepts; they are tools already enhancing productivity and resilience. Even a modest gain in efficiency, like 2%, can translate into hundreds of millions in value.

Abhishek and his team members focus on what he calls “micro-intervention thinking.” Rather than overhauling entire systems, they begin with targeted, measurable use cases. For example, Ask Emma is a Cargill generative AI tool that generated 140 new snack concepts in just a few days. Project Galleon opens in new tab/window uses AI to analyze poultry gut microbiota and provide data-driven recommendations for improved feed additives.

“We don’t start with a massive problem,” Abhishek says. “We start small, build what works and then scale.”

From creators to curators

As AI capabilities evolve, Abhishek sees a shift in how R&D professionals operate. Rather than building every solution from scratch, they’ll become curators of intelligent options.

“People use the term VUCA: volatility, uncertainty, complexity and ambiguity. These are all factors that make a situation harder to analyze,” Abhishek explains. “So, how do we actually make decisions and make it less about throwing a dart at the board and hoping it sticks? In the future, we will ask AI to generate thousands of possibilities — product ideas, microbiomes and materials. Then, we will use our expertise to select and refine the best options.”

Teams at Cargill are already seeing results. AI-powered platforms now recommend new ingredient combinations based on performance data, helping scientists reduce development time while increasing the likelihood of success.

“This tool learns from our recipes, lab experiments and their performances, and then suggests candidate recipes we should try in the bench lab experiments,” says Abhishek. This collaborative approach merges human judgment with machine intelligence, enabling researchers and operators to concentrate on what truly matters — solving the right problems.”

Balancing enthusiasm with pragmatism

Abhishek takes a measured approach to AI. “There’s plenty of hype,” he says. “The real challenge is integrating AI into how people actually work. There are those who are genuinely passionate about it and others who are quite paranoid about it. But the reality is always somewhere in the middle.”

He also acknowledges the challenges of implementing AI at scale. “AI is very different from past technological evolutions. People have figured out how to use it to boost personal productivity,” he shares. “But for AI to truly create business impact, the user experience needs to evolve. It must become intuitive and embedded — an extension of how teams naturally collaborate. And that's where the real value will be.”

Technology with purpose: multimodal AI

Abhishek is particularly energized by the potential of multimodal AI — systems that can interpret various types of data, from PID diagrams to plant layouts — and specialized models that integrate AI with domain-specific scientific knowledge.

“I’m excited about when general-purpose technology models get fused with very specific first-principles knowledge of climate, geospatial, physics, chemistry,” he says. “That’s when it becomes incredibly impactful. These hybrid models could accelerate the discovery of new materials and ingredients.”

But for Abhishek, the ultimate goal extends beyond technological advancement: It’s about using AI to help people make better decisions — for a better world. “We’re not just using AI to solve technical problems,” he says. “We’re using it to tackle real global challenges that impact people daily — like making food more accessible and helping the food system adapt to disruptions.”

That’s why he urges transparency, curiosity and hands-on exploration. “We shouldn’t be afraid of AI. Be tinkerers. Be open about how it’s used. And never forget the ethics that underpin good science – they still matter.”

In that way, AI becomes more than a tool for efficiency — it becomes what Abhishek has always sought: a way to connect the dots, reduce uncertainty and help people make the decisions that matter most. In a world facing unprecedented challenges, that clarity isn't just valuable — it’s essential.