22nd May 2019
How might we...
...use the limited design palette of conversational interfaces to save a visitor time on their next visit, while putting them in control?
Offering tailored experiences while minimising data requests
AI-driven services don’t need a huge amount of personal data to work — if we’re smart about context and the data we have, we can offer a compelling experience just by asking someone a few simple questions.
Product Context
Vouch participated in the TTC Labs Data Innovation Program, which was part of the first season of Startup Station Singapore in partnership with IMDA.
Vouch is an AI-powered chatbot concierge. Used by businesses like hotels, tourist attractions, and universities, their platform interprets and answers common guest and visitor questions via popular text chat platforms.
Vouch currently collects basic social profile information on people when they first interact with concierge chatbots powered by their service.
Problem & Opportunity
By remembering key details from previous visits and adjusting a guest’s profile accordingly, an opportunity exists here to offer repeat visitors or guests a more personalised and tailored experience at venues running Vouch’s platform.
At the same time, it is important to acknowledge that an “AI-powered concierge” will be a new concept for most people. To put them at ease and keep them informed, the service must clearly offer both transparency around what data is stored by the platform and control over which exactly which data points make up their visitor profile.
Team Vouch discuss concepts to develop
Design Features
Recommendations based on broad guest profile only
In this example, the Vouch bot is recommending activities that are popular with families.
By starting with such a broad categorisation, the chatbot can make tailored recommendations to guests without first asking them to engage with the service.
A “family” tag could be added to a guest profile on the platform based on a single question from front desk staff at check-in, or even inferred from which type of room they booked.
Design Features
Inferring details from context, rather than asking
Later in this conversation flow, the chatbot provides directions to a guest.
The service knows their destination from earlier in the conversation, and can infer a starting point from other data. For example, perhaps the guest started the conversation from the hotel’s guest Wifi network, or they asked for a hotel amenity (e.g. room service) immediately beforehand. At this point, guests are handed off to a third-party dedicated mapping service.
Next steps
The example conversation in this article uses the fact that a guest is travelling with their family as the starting point for a conversation.
What other easily accessible data points would be useful in the context of hospitality to create a personalised customer experience? How could something like this be translated to chatbots in other contexts — e.g. a shopping centre, public space, or tourist attraction?
