When I titled my recent conference talk “Are you being served?” one of my colleagues on LinkedIn promptly responded, “Mr. Humphries, are you free?” This is, of course, because I’ve surrounded myself with people who “get” BBC references and who are familiar with the 70s-80s sitcom involving the staff of a department store. In between hijinks, the shop assistants connected with their customers for a specific purpose: to lead them toward an informed and personalized purchase, while providing excellent CX along the way.
This kind of service is something that is often missing in the customer journey today. We are faced with increasingly large virtual and physical storefronts, but increasingly small available staff for assistance.
There are two different kinds of shoppers that need a higher level of assistance:
- The shopper who knows what she wants, but cannot find it.
- The shopper who knows his own tastes or needs, but not which product is best suited to them.
For most shoppers in these situations, we turn to research techniques that involve poring through product reviews, searching for answers to our questions online, and polling our friends on social media. But none of these techniques is reliable or efficient; reviews are often contradictory (1 star from one buyer, 5 stars from the next), as are answers to our questions; and our social media contacts may have experience, but it is most likely narrow. Entering terms into a search bar can be difficult when we don’t know which terms correspond to how the product is tagged in the back end database, when our search criteria are hazy (looking for something “small,” or “inexpensive”), or we don’t know enough about the products to even know what we are looking for.
What we need is someone to talk to, who can find out from us what we need or what we like, and who knows those products well enough to map this to a recommendation. This is an excellent space for a conversational virtual assistant to play a role.
The interaction with a virtual assistant can be as simple as taking the user through a few elicitation questions to find out what they are looking for. These questions can even map to the traversal of a decision tree, at the leaves of which are the recommendations. Or the questions could be like the “Which Hogwarts House do you Belong to?” online quizzes, seeking to score personality traits that tell you whether you should be wearing a scarf that is crimson, yellow, green, or blue.
In any case, conversation is the primary way humans learn about each other; it is natural that conversational technology should therefore be leveraged as a tool in this way. Virtual assistants can be designed to encode reliable product information and expert advice, while being available 24/7 in a webchat window, at a 1-800 number, or on a mobile device.
I am very aware that this may evoke the spectre of the expert systems of the 1980s. The critiques of those expert systems tend to focus on the fact that experts don’t always know what explicit factors are involved in their recommendations, making it a challenge to encode those decisions reliably and without bias. While this is true, creating a classification task for a more fashionable deep learning approach requires large data about shopper characteristics that we don’t necessarily have (not to mention the fact that deep learning has been just as subject to unintentional bias as any older approaches). If there is data, leverage it; the customer experience remains the same, with the ready virtual assistant asking how they might help the user, guiding them efficiently to the product, and creating brand loyalty through an excellent customer experience at the very entrance to the customer journey pipeline.
To learn more about how virtual assistants can support retail customers, check out our eBook: https://www.interactions.com/resources/industry/customer-experience-preferences-retail/