AI is mixing up how liquor brands make and sell their spirits

To pinpoint my favorite whiskey, I don’t ask the bartender at my local watering hole or the salesman at the liquor store around the corner. I turn to artificial intelligence.

What’s Your Whiskey is an AI-powered platform that was launched in 2019. It asks a series of questions to analyze and map a person’s flavor preferences against a database of foods and aromas. When I try it I’m asked, how do I feel about rosemary? Do I like chilies? I clicked on “Feed me fire!”—the most affirmative response. Ginger, orange juice, and apricot are among the other aspects of food I’m asked to weigh in on. Some flavors I love, while others evoke ambivalence or even distaste.

My unique taste profile is spicy, and my 99% match is the Glenkinchie 12, while Mortlach 12 gets 85%. Both are single-malt Scotch whiskeys that are unsurprisingly owned by liquor giant Diageo, which acquired Vivanda, the owner of What’s Your Whiskey, in 2022. What’s Your Whiskey is available in 29 markets and 18 languages today.

“It helps us connect consumers to the whiskey they might enjoy the most,” says Devin Nagy, Diageo’s director of technology and emerging platforms. “And you could see that translating to other categories we have in our portfolio like tequila and beer.”

AI has become an important ingredient for the liquor industry as it evolves sales and marketing tactics, alcohol production, packaging design, and how brands are built and engage with drinkers.

Among the AI tactics being deployed by Diageo include a paid social media buying tool called Smartbidder, which it codeveloped to ensure every dollar spent on media is as efficient as possible. Diageo is also using a generative AI tool to experiment with packaging designs for the company’s Smirnoff vodka and Don Julio tequila brands.

“I firmly believe that generative AI has the potential to change the way we approach productivity and creativity in marketing,” says Nagy.

Four years ago, Pernod Ricard began a comprehensive review of big digital investments that could potentially boost growth. Led by chief digital officer Pierre-Yves Calloc'h, the French beverage giant started with roughly 100 ideas and whittled them down to 19 to further explore the complexity and return on investment. Six of those projects were approved, four of which were powered by AI with a focus on sales and marketing.

With Pernod Ricard spending roughly 1.6 billion euros ($1.7 billion USD) on marketing annually, Calloc'h explains that optimizing those tasks with AI is extremely important. One of the projects now deployed across 13 countries uses AI to make store visits more efficient for Pernod Ricard’s sales representatives.

The sales team can have up to 1,000 stores in their purview, but they can only visit a few dozen each week. The mix of stores that could be visited each week could result in millions of unique combinations. In the past, larger stores were more frequently visited than smaller locations.

But today, data and AI help ensure store visits are smarter. Pernod Ricard looks at what each store sells, the nearby population, the size and assortment of the store, and other factors to create clusters of stores that should behave the same way based on that data. Sales representatives are then advised to visit the underperforming stores within those clusters where there would be potential to upsell. AI is also used to highlight whether a promotion is needed to support a Pernod Ricard brand like Absolut vodka or Glenlivet Scotch.

“The AI tool is giving precise recommendations of the five actions to be taken that day in the store,” says Calloc'h. “We are optimizing the time and impact of each sales rep on a day-to-day basis.”

In recent months, Pernod Ricard has asked employees across the company to submit their best ideas to use either generative or analytical AI. Calloc'h says 108 projects were collected, and nearly three dozen of those ideas will be reviewed more thoroughly to determine where the company wants to invest time and dollars.

“If you don’t put enough focus and enough money on [the use cases], they stay on the laptop and are not really used,” says Calloc'h. “The technology needs to be embedded in day-to-day processes in order to have an impact.”

His ultimate vision for AI is the following scenario: A shopper is on social media, perhaps while also planning to host a barbecue. If Pernod Ricard is doing everything right, that shopper gets an advertisement on their phone promoting a brand like Skrewball for a BBQ; then, when they go to their local store, the sales representative uses data to ensure the peanut butter whiskey is stocked at the right price. Consumers, Calloc'h says, must “know about the brand. And when they need it, they find it.”

“We are still in the very early stages of AI,” says Amit Parulekar, director of global advanced analytics and AI strategy at Brown-Forman, which makes Jack Daniel’s and Woodford Reserve whiskeys. “AI has lots of potential, but you have to be very careful about where and how to use AI.”

Brown-Forman utilizes a two-prong approach to AI. There are quick wins that help build exposure across the company to the benefits of AI, and longer-term initiatives that require commitment from senior leadership before investments are made. All use cases go through the company’s responsible-AI framework.

The company has built 200,000 elasticity models that are run monthly across the globe to determine how price increases or promotions would impact demand for Brown-Forman’s brands, as well as the predicted impact on competitor offerings. For whiskey production, AI models are predicting color accuracy for the brown spirits Brown-Forman produces. Machine learning models also help influence the production and projected demand for whiskey.

“What we are doing is carefully curating the use cases, accelerating and investing where necessary, but also piloting new technologies, such as generative AI,” says Parulekar.

“We are looking at potential applications of AI more from an efficiency perspective when we talk about assets and creation using our own imagery,” says Laila Mignoni, global head of brand marketing and communications at Bacardi. “AI is one ingredient, but it’s not the chef.”

Mignoni envisions AI as a tool that can help enhance the creative process, and the technology is just starting to be used in the company’s marketing events. One project involved giving five musical artists an opportunity to feed their musical tracks into a generative AI tool that was trained on the beats of Grammy Award–winning producer Boi-1da’s unreleased catalog. The result was an AI-powered EP that was released publicly in December.

“We’re interested to see how we can blend art and science,” says Natasha Curtin, global vice president of Bombay Sapphire, another brand owned by Bacardi. “How do we use it for data and insights that give us something? But how do we always make sure we always have that human skill, craft, and connection?”

Part of Bombay Sapphire’s “Saw This, Made This” installation, which was displayed in New York City, featured an AI-powered robot called Ai-Da Robot that painted artwork based on human submissions. Ai-Da Robot also created the artwork featured on two limited-edition Bombay Sapphire bottle labels. Proceeds from these AI offerings will be donated to emerging human artists.

Curtin says that is intentional, as the brand wants to support human-robot collaboration, rather than using technology to oust artists. “It is the combination of the two,” says Curtain. “Never seeking to replace.”

This story was originally featured on Fortune.com