Harvard Data Science Review Podcast

Episode: Better Data Science and AI Technologies for Better Vine and Wine?
Date: August 29, 2025
Host: Shao Lehman (Founding Editor-in-Chief, HDSR), guest co-host Magill Paredes
Guests:

• Katerina Axelson, Founder & CEO, Tastry
• Kia Pania, Co-founder & CEO, Scout and Neotempo Wines

Overview

This episode of the Harvard Data Science Review Podcast explores the intersection of data science, AI, and the wine industry. Through lively discussion with Katerina Axelson (Tastry) and Kia Pania (Scout, Neotempo Wines), the podcast investigates how cutting-edge technology is revolutionizing wine production, vine management, and consumer taste prediction. The participants delve into data methodologies, challenges unique to wine, the implications for jobs and industry efficiency, and the future possibilities for AI in agriculture and beyond.

Key Discussion Points & Insights

1. Guest Introductions and Company Overviews

[00:02-05:13]

• Kia Pania: Scout’s Vision

  • Kia's background in tech and love for wine led him to create Scout, a company using affordable camera technology and AI to create a detailed, actionable “medical record” for every grapevine.
  • Major insight: The best sensor in a vineyard is the farmer's eye, but it doesn't scale—so Scout scales this with AI and computer vision.
  • Quote:
    "We take 20 photos of every single plant... and build a database basically of what you have." —Kia Pania [04:11]

• Katerina Axelson: Tastry’s Approach

  • Katerina’s journey from chemist to founder; Tastry’s unique R&D led to AI that “tastes and smells” wine, predicting consumer preferences with high accuracy.
  • Tastry creates “digital twins” of both wine chemistry and consumers to simulate matches and identify market opportunities.
  • Quote:
    "We taught an AI how to taste and smell." —Katerina Axelson [06:51]

2. Data Collection and Ensuring Quality

[08:50-17:42]

• Scout’s Data Collection

  • Focus on surpassing human error; early discoveries showed that most clients' existing data was inaccurate.
  • Implementation of robust data quality checks, including a specialized data czar, risk scoring, and AI-driven blind spot detection.
  • Ongoing challenges include occluded fruit in photos and the necessity of occasional human calibration.
  • Quote:
    "In the beginning... I was very nervous because 100% of our data was different than the customers. And what we found... their data was wrong." —Kia Pania [10:57]

• Tastry’s Data Approach

  • All data is generated in-house: advanced chemistry analysis (via a proprietary, certified lab) and unique consumer sensory data.
  • Tastry triangulates chemical, sensory, and cultural preference datasets to model and anticipate future consumer choices—even creating synthetic data to model wider palates.
  • Quote:
    "What our data allows us to do is... anticipate the future." —Katerina Axelson [15:15]

3. Applications: Beyond Grapes and Wine

[17:42-27:11]

• Tastry’s Expansion:

  • The proprietary sensory technology is being applied to other verticals (e.g., coffee, soft drinks, fragrances), finding easier adoption and clearer value propositions in non-wine industries.
  • Wine is technologically challenging, with immense seasonal variation and unique, fragmented market dynamics.
  • Quote:
    “I realized how much easier it is to implement this technology in other industries so much faster...” —Katerina Axelson [20:30]

• Neotempo Wines and “Farming First”

  • Neotempo is a data-driven, sustainability-focused winery using electric tractors, precision irrigation, and open source sharing of smart farming tech.
  • Example: Data-driven irrigation saved vines during the 2022 heat wave. Use of Tastry allowed them to differentiate wines even from adjacent vines.
  • Quote:
    “Almost every function is data-driven and technology enabled.” —Kia Pania [22:27]
    “We make wines that are unique. And when you say that, they have to be unique even across a family.” —Kia Pania [24:29]

4. Data Science in Practice: Models and Infrastructure

[27:11-32:54]

• Scout:

  • Collects data as time-series to track vine changes over time.
  • Uses regression, anomaly detection, and “Vinayard Intelligent Assistant” (VIA) for integrating both private and public viticulture knowledge through LLM-based chatbots.
  • Emphasizes usage of proven, generic horizontal technologies combined with custom ag/viticulture models.
  • Quote:
    “The data set that says what has changed is also very, very important and in effect becomes a new data set.” —Kia Pania [28:12]

• Tastry:

  • Developed a custom, possibly liquid neural net-style, AI model for associating disparate data.
  • Incorporates Bayesian ridge regression and continuous model updates for market and palate predictions.
  • Quote:
    “We had to find a unique way to associate disparate data sets in a multidimensional space... that's really cool piece of our secret sauce.” —Katerina Axelson [31:02]

5. AI’s Opportunity and Risks in Wine and Beyond

[32:54-39:49]

• Opportunity

  • Wine is historically inefficient; AI can transform yield forecasting, resource usage, and profitability.
  • AI won’t just take jobs—it will eliminate tasks people don’t want (e.g., manual vine counting), streamline inefficiencies, and enable new, more interesting roles.
  • Quote:
    “It’s the only multibillion dollar business... that doesn’t have a forecasting model until the fruit is literally picked.” —Kia Pania [35:04]

• Risks and Human Role

  • Entry into wine was challenging; initial fear was that AI would replace winemakers, but in reality, AI-enabled rapid crisis mitigation (e.g. fire/smoke seasons), protecting jobs.
  • Human guidance remains essential; AI is a “paintbrush” for the industry’s artists, not a replacement.
  • Quote:
    “It’s really the human that is guiding and in charge of the AI. It is, it's your slave, not the other way around.” —Katerina Axelson [38:18]
    “It's really humans using AI that are replacing humans.” —Katerina Axelson [39:36]

6. Magic Wand: The Future of AI in Wine

[39:53-44:29]

• Tastry’s Wish: Expansion into new sensory-driven verticals. New AI to correlate consumer preferences across different product types—e.g., can preference for oaky chardonnay predict love of butter pecan ice cream?

  • Quote:
    “If I know that you like these oaky, buttery Chardonnays... does that mean I can predict that you're going to like the butter pecan ice cream?” —Katerina Axelson [41:43]

• Scout/Neotempo’s Wish: Fully integrated forecasting combining real-time vineyard conditions with weather and other time-series data, enabling proactive vineyard management on a whole new level.

  • Quote:
    “If I could have a magic wand... I would love to be able to get to a forecasting model that could combine forecasting and what we see on the ground with weather.” —Kia Pania [42:18]

Notable Quotes & Memorable Moments

• "The answer is easy. It's the farmer's eye. ...The only problem is that doesn't scale." —Kia Pania [03:28]
• "Our company tagline is we taught an AI how to taste and smell." —Katerina Axelson [06:51]
• "In the world of AI, you take that raw basic principle and start building models and ways that you can calibrate it. Because we don't know what the future holds. We know what the past holds." —Kia Pania [43:14]
• "It's really humans using AI that are replacing humans." —Katerina Axelson [39:36]

Timestamps for Key Segments

• 00:02-05:13 – Introductions & company overviews
• 05:35-08:50 – Tastry’s creation and unique approach
• 08:50-17:42 – Data collection and quality in Scout & Tastry
• 17:42-27:11 – AI applications beyond wine; Neotempo's smart farming
• 27:11-32:54 – Data analysis: models, infrastructure, continuous improvement
• 32:54-39:49 – AI’s impact on jobs, efficiency, and industry transformation
• 39:53-44:29 – “Magic Wand” wishes for the future of AI and data science in wine and beyond

Final Thoughts

This episode illuminates how AI and data science are reshaping even the most traditional of industries. Both Kia Pania and Katerina Axelson highlight the potential of technology to make wine production more sustainable, efficient, and consumer-focused—without losing the human artistry at its core. The future promises cross-industry sensory prediction, real-time yield forecasting, and deeper understanding of the connections between what we grow, create, and consume.