AskJerry - Case Study
Learnings from WeChat's first A.I. powered bartender
Jan 2, 2018
Chinese millennials don’t drink the same liquor that their parents did. The centuries-old custom of guzzling baijiu during business dinners has been replaced with the millennial preference for sipping cocktails with friends in a trendy jazz bar. China’s per capita alcohol consumption is on the rise. But the growth in consumption is uneven among different types of alcohol. Millennials are much more likely to express their own personal identity in various kinds of cocktails and wine.
Even among millennials, cocktail culture is far from mature. Millennials depend on trusted online peer reviews, apps or clubs that allow them to discover new drinks and indulge their aspirations, as exotic drinks pop up in the nation’s bars for the first time.
Tomorrow China, a Shanghai based agency, approached Pernod Ricard with their version of the consumer insight: if users chatted with a trusted cocktail expert, they were more likely to consume more cocktails in the future. To do this, Tomorrow China proposed personalizing cocktail conversation with China’s burgeoning post 1990s demographic in order to build awareness about cocktail culture and insights.
To achieve this however, Tomorrow quickly ran into a scalability issue. They were not able to track user insights, and engage with users in a scalable way due to the large volume of demand. However, the company needed a way to track consumption data and keep customers coming back. The company needed to expand its customer service capability to keep up with the consumer queries coming in. In short, Tomorrow needs to:
- Pursue personal connections with users in an efficient SCALABLE manner.
- Track user behavior and build profiles from these connections to increase sales.
However to make these insights actionable, Tomorrow China, needed to use innovating A.I. technology to make this scalable and cost effective.
Rikai Labs blends chatbots powered by Artificial Intelligence, Natural Language Processing and machine learning with live people to engage and educate WeChats 900 million users.
Together, Tomorrow China and Rikai Labs built the Ask Jerry official public Wechat account. Rikai Labs’ technology allowed Tomorrow China to chat with Ask Jerry’s followers at scale while also tracking user data to gain customer insights and more effectively sell products to users.
Jerry is a suave, sophisticated bartender who offers cocktail recipes and suggestions tailored to the individual user. He recommends drinks and promotes weekly cocktail articles, and provides links where customers can buy cocktail ingredients from his shop. After chatting with Jerry, users can purchase Pernod Ricard cocktail kits with everything that the user needs to make their own cocktails at home.
As Jerry chats with users, it’s N.L.P. engine collects user data and segments users in real time to provide links to specific products. Rikai Labs’ chatbot engine provides back-end analytics to track demographic behavior to consistently provide drink recommendations.
How it works
Users engage with the Ask Jerry account using a blended chatbot and customer support engine experience. The typical user journey consists of:
1. Educating Users
Based on users preferences, the Ask Jerry account will then offer relevant information, from how to make the perfect cocktail to other article recommendations.
2. Collecting Data
When users follow the official account, Jerry welcomes the user and asks them about their tastes in cocktails. The conversation runs through Rikai’s Natural Language Processing engine, allows Jerry to build up a database of user information and make tailored recommendations.
3. Converting Users
When appropriate, the account provides links to the e-commerce store that that are relevant the users have been personalized based upon the users flavor profile and interests.
By implementing Rikai Labs' chatbot platform, Tomorrow China were able to solve problems in terms of e-commerce
- Engagement at scale
- Increase conversions
Aside from driving sales, Jerry is able to better educate users about cocktail culture. Through the course of offering tutorials on how to mix up the perfect Piña Colada, Jerry will also ask about the user’s alcohol preferences, and make better recommendations based on the user’s profile. This personalized approach allows for more accurate data insights. By automating the customer engagement system, AskJerry also makes it easier for customer service to engage and track the large amounts of consumer data generated from consumer interactions and sales.
- 79% of users monthly active users engage Jerry for cocktail recipes.
- 375 customers chatting with only 1 person from Tomorrow Staff
- 5.8% of total messages (User, Agent, Chatbot) were sent by a live agent
- 19.6% month on month user acqusition
- 5% increase in sales for every 6 chats with ChatBot
By blending A.I. powered chatbots with live customer support agents, Tomorrow China was able to open up a new distribution channel for Pernod Ricard that, absent technology, would have been too costly to operate. The chatbot not only helped to increase sales and revenue, but it also helped to automate and make internal work flows more efficient and automated. By blending the A.I. chatbot with live support, Rikai Labs helped Tomorrow China to make online sales more personal, efficient, and scalable.