13 December 2019
AI chatbot 'Antony' served as a virtual concierge during the Retail Asia Expo in 2016. Photo:
AI chatbot 'Antony' served as a virtual concierge during the Retail Asia Expo in 2016. Photo:

Chatbots: ‘Say hello to my little friend’

Need some information quickly? Chances are if you are online typing in your request, a chatbot will be dealing with it. The industry is developing in leaps and bounds and becoming essential for customer service in retail, particularly with the volume of customer interactions occurring online. Eugene Lai, Tech and Transformation Manager for Shiseido Group Hong Kong, knows all about the challenges of creating a chatbot, having developed ‘Antony’ in a previous role for Red Ant Asia. 

Excerpts from a conversation:

Q: First, what is your experience in creating chatbots?

A: At my previous company, Red Ant Asia, I led a team that developed Antony, an AI chatbot as a proof of concept for the Retail Asia Expo back in 2016, where he served as a virtual concierge, giving details about food and beverages options, key locations around the expo, real time weather and background information on Red Ant.

The majority of AI chatbots use a messenger-based communication tool between AI and customer, but with Antony, I wanted to take him a step further and introduce verbal communication as a way of engaging with him – a method of bridging the gap between virtual and real world interaction; by essentially replacing touch interaction with verbal interaction.

Q: The process of helping chatbots learn is very hands on, can you explain how it’s done?

A: Every chatbot must have a predefined, use case, such as a customer service chatbot. From there I build a list of all possible questions that a customer would possibly ask and the different ways of asking. Examples might be: “what is your office located?” or “can you tell me your office address?”.

Then I’ll write out the answers to these questions and to give the chatbot a little more variation or ‘personality’, I’d write out a few different ways of answering a question. The more effort that is put into this initial period, the more ‘knowledge’ your bot will have. It’s important to note that the bot needs to continuously learn.

Q: Can you provide a personality to a chatbot?

A: Many chatbots that I’ve seen lack any personality and are often just really boring to interact with so creating a personality for your bot is a MUST for me.

Part of the fun of developing an AI chatbot is the ability to define their characters through what words and nuances we train it to use in response to a question. While the developers worked on the nuts and bolts of Antony, I worked on his speech responses, so he ended up having a lexicon similar to mine. I would treat this as a creative task as if you were writing the script of a character in a movie.

Q: How do you account for wildcards in language?

A: Through training, most chatbot platforms have an array of “I don’t know” answers which can be used to address when they don’t understand a user input. In addition, most chatbot platforms have a training backend where you can view all questions asked and their response. For all questions that the bot didn’t know how to answer, you can either a) assign that question to an existing answer (also called an intent), or create a new answer.

Q: How do you see the industry developing?

A: I guess in terms of chatbots, the areas of Artificial Intelligence that drive it, will advance and we’ll become more reliant on using voice to interact with computers. The success of home assistants like Google Home and Amazon Alexa is indicative of people becoming more comfortable with voice commands. I’m all for voice commands as I’ve built my own home assistant leveraging off on Amazon Alexa, that allows me to do the standard home automation (turning on and off lights, A/Cs, etc), and also get greeted when I get home, or get notified on my phone if someone has come into my apartment.

– Contact us at [email protected]


EJI contributor