How data and AI are helping Deschutes Brewery respond to beer trends faster
Transform recently sat down with several Microsoft customers at an event highlighting emerging trends in data and artificial intelligence (AI). We spoke with Kyle Kotaich, operations technology lead at Deschutes Brewery, a craft beer maker based in Bend, Oregon, that combines data science with brewing expertise to deliver better beer and run a more efficient business.

TRANSFORM: How does AI help you run a better business?

KYLE KOTAICH: It takes a little while for us to get a new beer from, say, pub scale or research and development, all the way to production scale when a trend is still hot.

So, one of the things that we talk about amongst the team is, ’Are we able to use data and AI to spot trends, maybe in other parts of the country or even other parts of the world and make a correlation and predict what a trend might be?’ And that way we can be less reactive and more progressive.

TRANSFORM: Are the people outside your technology team using data and AI?

KOTAICH: We have a lot of support from other departments. Within our brewing quality and packaging group, [which handles] the actual manufacturing of the product, people are very comfortable with the concept of what we’re doing and how we’re using machine learning to become more efficient. Getting people to think about the possibilities is the first step.

TRANSFORM: Is the technology around data and AI improving?

KOTAICH: I definitely think so. Our goal is to increase operator and business awareness just by using these tools and making them very available — whether that’s real time data, predictive analytics or just access to raw data.

Our Azure SQL Database has been really great for combining a bunch of data. And once it’s there, there’s so many options. We use Azure Machine Learning Studio to do machine experiments. We use the Azure SQL Database to send text alerts to contractors that come in to pick up our brewery waste, when the tanks are full, and compost it. We use it for a number of different things.

TRANSFORM: Are the tools themselves getting easier to use for people that don’t necessarily have deep AI or data science knowledge?

KOTAICH: I can use my personal experience. I didn’t have any prior knowledge of AI or machine learning or data science other than my background in physics. And becoming a brewer exposed me to the value of good data for quality and efficiency.

I was able to learn the data science part of our projects really quickly, because of the tools that are available, whether that’s Azure Machine Learning, Jupyter Notebook, or all these different platforms that are available. They are very well documented and easy to use. It makes these projects feasible for anyone.

TRANSFORM: Where is Deschutes on its AI journey?

KOTAICH: There’s definitely a lot of room for growth. The projects that we’ve [undertaken], we’re still fine tuning them and we always will be. And I think there are really good cases to say we should continue doing these types of projects.

We have a brewery operations technology team. It’s a relatively small team, five people. And we’re looking for more ideas of where we can further streamline our processes.