AI & Wine: Talking Tech with Prof. Dr. Dominik Durner
First demonstration of how the principle of the project works at the ProWein fair in Düsseldorf in 2023. Photo: Wille Engineering
I got the chance to talk to with Prof. Dr. Dominik Durner, an expert in wine science, currently serving as Professor of Food Technology & Enology at the Weincampus Neustadt in Germany. His work explores the intersection of winemaking and technology, focusing on AI-driven innovations such as the PINOT project (an artificial nose for wine) and SmartGrape (AI-based grape quality assessment).
Artificial intelligence in winemaking is a topic that fascinates me, not just as a sommelier but as someone who believes in the balance between tradition and innovation. Technology has always played a role in shaping wine, from the invention of the wine press to temperature-controlled fermentation. AI is simply the next step in this ongoing evolution.
AI is not just a tool for efficiency, it will over time mean better quality, sustainability, and consistency in wine production. Ultimately, it will shape the wines we, as sommeliers, work with and recommend.
When I came across the SmartGrape and PINOT projects, I knew I had to learn more. That’s what led me to speak with Prof. Dr. Dominik Durner.
"Hello, and it's my pleasure to be here today!" he greeted me. "I'm excited to talk about artificial intelligence in winemaking, it’s a fascinating field to research and explore."
“It’s a rapid change, and anyone who thinks AI won’t impact viticulture and winemaking is wrong. It absolutely will.”
The Journey into Wine Science
To start, I asked him about his background and how he became involved in applying AI and technology to wine.
Prof. Dr. Dominik Durner
Professor for Enology and Food Technology, Univ. Appl. Sci. Kaiserslautern, Weincampus Neustadt
Photo: Stephan Presser photography
"I’ve been a professor in food technology and enology for 15 years now. These two fields are much closer than one might think. Winemaking has always been a field of innovation, if we look back hundreds or even thousands of years, both the process and the wine itself have changed. The way wine is made influences its style, and technology plays a key role in that transformation."
He continued:
"From an innovation perspective, there are two sides to this. On one side, we want to push quality and explore new wine styles—such as wines with lower or no alcohol that still taste like traditional wines. On the other hand, technology should not take away from the craftsmanship of winemaking. Every wine is unique, and technology should support, not replace, human creativity."
AI in Winemaking: Revolution or Evolution?
I asked him whether AI represents a revolution or an evolution in viticulture and vinification.
"It’s both, a revolution and an evolution at the same time!" he said with enthusiasm. "It’s a rapid change, and anyone who thinks AI won’t impact viticulture and winemaking is wrong. It absolutely will. But at the same time, winemaking has always evolved. In my view, AI will follow a similar path. It will enhance quality, support authenticity, and provide winemakers with better insights, rather than replacing human craftsmanship."
The SmartGrape Project: AI in the Vineyard
Prof. Durner shared details about SmartGrape, an AI-driven system designed to assess grape quality.
"When it comes to harvest, we all know that grapes are assessed by a refractometer to see the sugar content, the total solutes and this is not enough. I mean, we want and need to assess the grapes in terms of how the future wine could look like. Artificial intelligence and machine learning especially have a lot to do with forecasting."
He continued, illustrating the potential of AI in viticulture:
"So imagine we have a tool, and you look through that tool and do your assessment, not only seeing the sugar graduation but actually seeing a forecast of how the wine could be just by looking at the grape."
The project focuses on Riesling grapes as a test model:
"Yes, we are talking about a lab environment, but the idea is to see (like in our case, mostly Riesling) how these grapes can already tell us what the wine is going to taste like. And this is absolutely fascinating, what is possible nowadays with analytical tools combined with computer models, translating analytical signals into something more enological, making a forecast."
The next step, he explained, is bridging the gap between data and decision-making:
"Then it’s about making the AI talk to you, like we have chatbots now, right? So imagine a system telling you, 'This wine is going to be like this and this.' That’s where we are headed."
At its current stage, SmartGrape is a handheld device with a diamond-based ATR sensor:
With SmartGrape, we can instantly predict sugar, acidity, and nitrogen levels. This allows winemakers to make informed decisions about harvest timing. You press a grape berry onto the sensor, and it provides an analytical reading. The challenge is that for now, it still requires manual sampling—we're working on making it scalable."
"In the future, we might see this technology mounted on drones or vineyard machines, automating the analysis and giving real-time feedback on grape ripeness and composition. That would be a game-changer for vineyard management."
Research prototype laboratory version of the PINOT project for the automatic determination of various wine parameters such as headspace/gaseous analyses, temperature, conductivity, pH value and photometric value. Photo: Wille Engineering
The PINOT Project: The Artificial Nose
One of the most exciting projects (in my view) Prof. Durner is working on is the PINOT project, an AI-powered artificial nose that detects wine aromas and faults.
"We’ve trained the AI to recognize certain faults earlier than a human nose can. For example, we can detect sulfur-related off-aromas before they become a problem."
When I asked him about the challenges they faced while developing the artificial nose, he was candid:
"Ethanol was a big issue. It overwhelmed our sensors at first. We had to refine the system to focus on specific volatile compounds, and that took time. Another challenge is the data, it’s not easy to collect thousands of wine aroma profiles, but we’re working on it."
The Challenge of Detecting TCA and Oak Influence
TCA (cork taint) detection remains a difficult problem for AI.
"We can train the sensors to detect TCA in a non-oaked wine fairly easily. But as soon as a wine has been oaked, the barrel aromas interfere with the sensors, leading to false positives. That’s one of the reasons we’re not commercial yet, he smiles.
Automating Repetitive Tasks
In the cellar, AI-driven tools are becoming increasingly relevant.
"Fermentation monitoring is one of the most important areas where AI can help. Sensors can continuously analyze CO2 production, temperature, and volatile compounds, alerting winemakers if something goes wrong."
But AI will also helping reduce manual labor.
"There are so many repetitive tasks in the cellar: monitoring temperatures, pumping wine, doing documentation. If AI and robotics can handle these, winemakers can focus on the real creative decisions: when to rack, how much lees contact to allow, and how to manage SO2 additions."
The Future of AI in Wine
Looking ahead, Prof. Durner is confident AI will play an even larger role.
"In five to ten years, AI-driven aroma detectors and vineyard sensors will be standard tools. We’re already working on making the artificial nose more compact, ideally the size of a smartphone."
He also believes AI will play a role in wine authenticity verification.
"We can create digital fingerprints of wines. Imagine being able to verify a wine’s authenticity instantly by comparing its spectral data to a certified profile."
Photo: Stephan Presser photography
“AI can analyze and predict, but deciding what kind of wine to make? That will always be up to the winemaker.””
Photo: Stephan Presser photography
At the end of our conversation, I asked him about the biggest hurdles to AI adoption in winemaking.
"Cost is a big barrier, especially for smaller wineries. Technology is usually an advantage for big producers first. But if AI is used correctly, it can help small wineries too, by improving customer outreach, optimizing production, and even aiding sustainability."
When I asked him what will always require human touch, he smiled and said:
"Creativity. AI can analyze and predict, but deciding what kind of wine to make? That will always be up to the winemaker."
As technology evolves, so too must our understanding of it.
AI is set to shape the wines we drink and recommend in ways we are only beginning to grasp.
