He also regularly finds himself among the brewing kettles. As an enthusiastic home brewer with his own Brewery Herman — named after cabaret artist Herman Finkers — he combines technology and craftsmanship. His setup is equipped with sensors and Raspberry Pis and can be controlled remotely. “My brewery is full of measuring equipment,” he says. “That naturally makes you wonder: what else can data add here?”
From data to beer
That question led to an experiment. Berry combined open-source recipes from BrewDog with data from the beer app Untappd, where users worldwide rate beers. “The idea was: can we train a model on recipes and ratings, and then predict what people like?” he explains.
The algorithm generated a new recipe based on popular characteristics. But it turned out to be less straightforward than expected. “An IPA is hoppy, a stout is dark and heavy. You can’t just mix those flavours.” In the end, a brewer had to adjust the recipe. “That immediately shows: domain expertise remains essential.”
More than just a smart recipe
According to Berry, the real strength of AI is not in creating the ‘perfect beer’, but in improving the entire beer value chain. “Recipe development is the low-hanging fruit,” he says. “Much more interesting is: how can you use data to make better decisions in production, sales and product selection?”
For this, he collaborates with companies including Untappd, whose European headquarters is based in Enschede. The app contains data on what people drink, where and when — valuable insights for breweries and hospitality businesses.
What goes on tap?
That data can help answer questions that entrepreneurs often rely on intuition for today. For example: which beers should you put on tap? And how often should you rotate them? “If you offer ten IPAs, but your customers only need one, you’re missing opportunities,” says Berry. “With data, you can better align your offer with what people actually want.” Preferences also vary by location. “What works in a student city might not work somewhere else at all. Data allows you to tailor your offering much more precisely to local tastes.”
Responding to a changing market
The beer market is changing rapidly. People are drinking less alcohol, and alternatives such as kombucha and seltzers are gaining popularity. At the same time, demand for locally and sustainably produced beer is growing.
“Organic malt is more expensive,” Berry explains. “But data might show that in certain regions there’s actually more demand for it. That allows brewers to respond strategically.” According to him, AI can help identify such trends earlier and support businesses in a competitive market.
From keg to flavour
There are also opportunities within the brewing process itself, for example in quality control. “I recently heard that beer from a keg can lose quality over time,” says Berry. “With data analysis, you might be able to predict when that happens, and prevent it.”
Smaller breweries in particular can benefit from this. “They often don’t have large R&D departments. Technology can help them work more intelligently and consistently.”
Technology that works in practice
As an industrial engineer, Berry operates at the intersection of technology and practice. His goal is not only to develop new tools, but to ensure they are usable for businesses. “It shouldn’t be a tech push,” he says. “You need to focus on where it actually adds value for entrepreneurs.”
He sees parallels with his own brewing project. Many breweries start as a hobby, but not all succeed in becoming sustainable businesses. “Data can help here too: supporting better decisions so brewers can remain economically viable.”
The future of beer and AI
For Berry, the main goal is clear: understanding where AI truly adds value in the beer sector. “And perhaps even more importantly: ensuring the sector continues to exist. I hope that in ten or twenty years we’ll still have vibrant squares full of cafés and craft beer. That depends on how healthy the entire chain is.”
