Harvard Data Science Review Podcast Better Data Science and AI Technologies for Better Vine and Wine?
Aug 29, 2025
Join Kia Behnia, CEO of Scout and a vineyard proprietor, alongside Katerina Axelsson, founder of Tastry and expert in sensory sciences. They delve into how AI is revolutionizing viticulture and consumer preferences in wine. Kia discusses the power of camera-based monitoring for grapes, while Katerina reveals her innovative approach to predicting wine likability using analytical chemistry and machine learning. Plus, they explore the implications of these technologies for sustainability and beyond, blending art and science in winemaking.
AI Snips
Chapters
Transcript
Episode notes
Farmer's Eye Scaled With Cameras
- Kia Behnia turned vineyard scouting into Scout by using cheap cameras and AI to create a medical record for each vine.
- They take 20 photos per plant to measure canopy, clusters, and shoots via computer vision.
Teaching A Computer To Taste
- Katerina Axelsson built Tastry by inventing an analytical chemistry method that captures complex flavor analytes together.
- They pair that chemistry with consumer panels to train AI to predict likability with about 93% accuracy.
Treat Data Quality As First-Class Metric
- Data quality dominates downstream performance and must be engineered into the pipeline as a metric.
- Scout assigns risk scores to photos and has a dedicated data-quality role to flag low-quality inputs.