Artificial intelligence has become one of the brewing industry’s most talked-about topics, but for many brewery owners, the conversation often swings between inflated promises and outright skepticism. Michael Perozzo, co-founder and CMO of Vice Beer in Vancouver, Washington, believes the most valuable use of AI sits somewhere in the middle.
Rather than asking artificial intelligence to design can labels, write social media posts or develop recipes, Perozzo has turned it into something much more practical: a business analyst that helps him better understand the financial and operational health of his brewery.
“I was just having one of those moments of, ‘What am I missing here? What am I leaving on the table? Where are my shortcomings?'” Perozzo said in a recent BREWER Podcast.
Like many owners of small breweries, he wasn’t looking to replace expertise. He was looking for another perspective.
With Vice Beer producing roughly 900 barrels annually, hiring a full-time chief financial officer or bringing in consultants on a regular basis isn’t realistic. Instead, he uploaded monthly profit-and-loss statements into ChatGPT and asked it to review three months of financial data for trends.
“I was absolutely blown away by what it came back with and the level of detail and spotting trends and pointing things out that I didn’t even notice,” he said.
For Perozzo, the biggest value wasn’t that AI generated answers. It found patterns he hadn’t considered after reviewing the numbers himself multiple times.
That experience reshaped how he approaches the technology. Instead of treating AI like an all-knowing decision-maker, he now assigns it a specific role.
“I often ask it to take on the role of my CFO,” he said. “I want it to give me that accountability and tell me, ‘You’re spending too much in this area,’ or, ‘Your payroll in this area, you need to make some decisions here.'”
The quality of those conversations, he said, depends heavily on how specific the prompts become. As he has continued using AI over several months, he’s refined those prompts to generate consistent reports and analyses that matter to his business rather than generic summaries.
The approach has also expanded beyond financial statements. Perozzo now feeds the system wholesale invoices, production numbers and retail sales information to help identify trends in sales channels, monitor key performance indicators and evaluate production decisions.
One example involved deciding whether to brew a Saison.
The beer made sense from a branding standpoint after adding a brewer known for producing the style. But AI highlighted an operational issue that hadn’t entered the discussion.
“It pointed out that seven barrels of Saison would tie up 15 kegs for five to six months,” Perozzo said.
The beer itself wasn’t the problem. The opportunity cost was. That kind of thinking has become one of the technology’s biggest advantages, Perozzo noted.
“It’s changed the way I think in general,” he said.
Perozzo said AI has become another voice during internal discussions, sometimes surfacing considerations that neither he nor his brewing team had raised.
“Often it’s something none of the three of us have thought about yet,” he said.
Those insights have influenced pricing discussions, sales goals and monthly KPI tracking, but Perozzo stresses that every recommendation still requires human judgment.
The software analyzes data exceptionally well. It doesn’t understand every local market.
“There are definitely still lenses you have to look at the data through … your own understanding of your market, your own practices and what you’ve found to be successful,” he said. “You can’t just rely on that.”
That distinction becomes especially important when AI recommends pricing changes or sales strategies based on national industry data.
While those recommendations may reflect larger market trends, Perozzo compares them to reading Brewers Association reports. National averages don’t always reflect what’s happening in an individual market.
“What would probably work in Oklahoma isn’t going to fly in Portland, Oregon,” he said.
That philosophy also shapes where he refuses to use artificial intelligence.
“I cringe at the thought of using it for recipe development,” Perozzo said. He feels similarly about relying on AI for major branding and design work. “There’s still a human element to art that AI cannot pick up on.”
Instead, he sees AI’s greatest strength in processing information quickly, identifying trends and helping owners ask better questions.
“It’s a supercomputer for mining data and giving you feedback,” he said.
He has even used it outside brewery operations, including reviewing a lengthy operating agreement before meeting with an attorney. By asking AI to identify relevant clauses and suggest questions ahead of the meeting, he was able to have a more focused legal discussion.
READ MORE: iBrewer: Sugar Creek Mixes Tech Into Brewing Process
“It made me more efficient in that meeting,” he said.
For brewery owners curious about AI but unsure where to begin, Perozzo doesn’t recommend experimenting with marketing copy or image generation.
“If you’re just hearing this and thinking, ‘Where do I start?’ I’d start with your P&Ls,” he said on the podcast.
That advice reflects the broader lesson from his experience. AI hasn’t become Vice Beer’s CEO. It hasn’t replaced management, brewing expertise or creative thinking. Instead, it has become something many small breweries have never been able to afford: a junior analyst that reviews the numbers, asks difficult questions and gives ownership another perspective before the real decisions are made.
