Susanna works to ensure that Hong Kong and China Gas Company Ltd., where she is the General Manager—Corporate IT, has no communication barriers with its 1.9 million customers. The company, known to customers by the nickname “Towngas,” receives about 5,000 customer inquiries each day. Susanna oversaw the implementation of an IBM Watson Assistant-based virtual agent called Tinny, which aims at handling 20% of general inquiries with 80% accuracy—ensuring customer satisfaction and easing the workloads of busy client service officers.

What has your company achieved with Watson?
Tinny now handles 10% of the general inquiries we receive on the company website, mobile app and WeChat. General inquiries represent 20% of the total customer inquiries in a day. Tinny provides text-based response in English and Chinese, including colloquial Cantonese, with an 80% accuracy rate.

What other measurable gains have you seen, and what do you hope to see in the future?
Tinny performed well on post-chat surveys, scoring an average of seven out of 10. Just as importantly, our call abandonment rate dropped from 3% in 2018 to 1% in 2019, when we rolled out Tinny. This year, we’re integrating Tinny with our live agents via e-chats.

What makes an AI project successful?
You first need to have a sound business case for how AI can address pain points and help your business growth. You also need complete buy-in from the person charged with overseeing the business result. Lastly, you have to be able to redesign processes throughout the transformation, to achieve a higher level of positive results. 

What do you know now about AI that you wish you had known when you started?
We started the dialogue around design of the chatbot with open-ended questions, but this approach increased the chances of Tinny incorrectly answering customer queries. It also entailed a lot of human effort to maintain the dialogues. Although the chatbot still allows open-ended questions, we re-designed it to always show a menu of options. By narrowing down the questions, we have improved the accuracy of our responses.

Why should women be more involved in the creation and use of AI?
Male leaders usually put more focus on technical achievement, while women leaders are usually more empathetic and collaborative. The qualities of a female leader offer a better chance of success in AI projects that need to involve different stakeholders. With more empathy, women have more tendency to put themselves in the shoes of customers and can deliver products that are more accepted in the market. 

She puts Watson Assistant to work

Dive deeper into your journey to AI and learn more about Watson Assistant