Analytical Flavor Systems

Applied AI
Enterprise

The Science of Predicting Taste

Story

The Limitations of Food Testing

The global food and beverage industry launches tens of thousands of new products every year and the overwhelming majority fail. Product development has historically relied on slow, expensive human taste panels—small groups of testers whose feedback is subjective, inconsistent across cultures, and difficult to extrapolate to new markets. For example, a flavor that wins in a Chicago focus group tells a brand almost nothing about how it will perform in Jakarta or São Paulo, leaving billions in R&D spent wastefully on guesswork.

Analytical Flavor Systems: The Science of Predicting Taste

We backed Analytical Flavor Systems because founder Jason Cohen saw that human sensory perception itself could be modeled—and predicted—with machine learning.

The company built Gastrograph AI, a predictive sensory intelligence platform powered by an enormous database of human flavor, aroma, and texture perception. Instead of relying on panels alone, Gastrograph's models could simulate how distinct consumer populations—segmented by demographics, geography, and preference archetypes—would perceive a product before it hit a shelf. The platform let major food and beverage conglomerates develop, test, and reformulate products, dramatically reducing development cycles and improving the odds of a successful launch.

We invested in Analytical Flavor Systems because it was building a foundational data asset in a new category: a proprietary, ever-growing map of human taste. It was particularly impactful in an industry defined by high failure rates and slow iteration.

Exit Journey

Gastrograph AI grew to support some of the largest consumer brands in the world, and its unique sensory database made it a strategic prize. In April 2025, the company was acquired by NielsenIQ, the global leader in consumer intelligence, to anchor the company's AI-driven product innovation capabilities. Analytical Flavor Systems validates a core Hyperplane target: proprietary data plus applied machine intelligence in an overlooked vertical which creates an attractive, acquirable, sustainable business.