Intuitively, it was only reasonable to select one problem and create a feature focussed purely on tackling it. However, in the grand scheme of things, our team wanted to create a feature that tackled most of the problems we identified, or at least create a platform with which solutions could be added on like building blocks over time. Hence, we wanted to revamp the existing Shopee chatbot (found in “Me” tab -> Help Center ->Live Chat) into a chatbot that was several consultants in one.
Not only would we be solving the problems of the users. Our solution would transform the current chatbot from a customer service tool into a sales and marketing tool that helps achieve business goals. Other than user specific needs, the chatbot could also promote bundle offers and sales when there are multiple items in the cart. Essentially, the chatbot is a feature that could enhance the sales process.
The scalability of such a solution would come in rolling out different consultants for different users. By tackling each user group separately, the experience of these user segments would not be side-lined. Separate teams could design personalized experiences for each segment, and training for each machine learning model could be separated.
Our inspiration came from a popular rise in chatbots in 2020 such as busuncle.sg and Facebook Messenger chatbots.
The first consultant we needed was for general enquiries, and any other edge cases not covered by other consultants. Sophie was already carrying out this task.
Next, we needed to expand on the functionality of the chatbot by including other consultants. To decide the types of consultants needed, we look to our findings from our user interviews and discovered the two main groups of users were most interested in clothing, and tech gadgets respectively. We could add two more mascots as the fashion consultant and technology consultant. Every consultant will be accessible from the chatbot interface as well as relevant listings.
Fashion Consultant (Foxy the Fashionista)
We will use the Fashion consultant to solve the problems of not getting the right recommendations and inaccurate sizing. To better enhance recommendations, the chatbot could be used to acquire crucial information such as body type, skin tone, and favourite fashion styles. When looking at a listing, whether recommended or not, the chatbot could also introduce an AR filter to let users determine how they look in it, and whether the sizing is correct. By sharing pictures of the AR filter with others, the user could also receive discounts to encourage word-of-mouth referral and generate growth.
The chatbot can also function as a feedback loop. For instance, the chatbot could jump in when a user rates a product poorly. The seller can offer discounts to stem disappointment and create questionnaires to understand and better serve their customer in the future.
Technology Consultant (Sean the Technologist)
We will use the Technology consultant to solve the problems of authenticity and provide quality of life features such as recommendations based on similar specification. In terms of authenticity, the consultant could provide more information on the seller, whether they are the official supplier or distributor. Current features such as Shopee Guarantee are in place to protect the buyer but does not provide sufficient transparency on the listing.
