With more people hosting smaller house gatherings instead of big, flashy parties, Diageo developed SIMI, the AI-powered bartender, to deliver an engaging, helpful brand experience in India.
Case study summary
• Targeted Indian city dwellers ages 25 and up
• Created AI-powered chatbot with 300 DIY recipes to make cocktails
• The bot became smarter with more use with more personalised recommendations.
• 3.2 million users interacted with the chat bot, driving 32% interaction rate
The challenge
In the past few years, the trend of personalization affected consumer behavior in a myriad of ways, including how people prefer to spend a night out. Large, flashy parties had increasingly become smaller, more private (and personalized) house parties. Diageo wanted to tap into the trend ahead of its competition, and deliver an experience that connected the brand to house parties organically.
Diageo’s target audience was city dwellers ages 25 and up, with focus markets in the eight largest metropolitan areas in India. This group was constantly on the move and always on their smartphones.
The solution
Based on Google trends and Facebook audience insights, the brand confirmed that its target audience preferred house parties to other forms of nightlife, and planned these parties over the web, specifically via mobile. Search trend analysis showed that innovative, personalized cocktails were a key to a successful house party, and consumers needed guidance crafting the perfect drink. Diageo created a chatbot through Facebook Messenger that helped consumers choose and mix personalized cocktails for their next event.
SIMI, the AI-powered bartender, was the India’s first branded chatbot, armed with more than 300 DIY recipes to make cocktails that consumers could easily put together using household ingredients. The more the bot interacted with consumers, the better it became at offering smarter, more personalized recommendations.
One of the key benefits of using a chatbot through Facbook Messenger was that consumers didn’t need to download an additional app to interact with SIMI. Interaction with the bot was simple, and the interface was something consumers were already familiar and comfortable with.
The brand targeted consumers exclusively through Facebook.
Results
The SIMI chatbot generated significant positive results:
• 10 million users were reached within first two months of the chabot going live on Facebook
• 3.2 million users interacted with the chat bot, driving an unprecedented 32 percent interaction rate
• 900,000 users requested recipes, with an average of two recipes requested per interaction
• More than 43 percent of the users returned to interact with the SIMI chatbot
