The team views the chatbot as guided discovery--the questions start big and then narrow down until the task is completed and the best recommendation can be given. In order to provide the best recommendation, the bot must understand what the traveler is really looking for--after all, a restaurant recommendation in Miami will be different based on if someone is traveling for work or pleasure. Chatbots have potential to be a powerful, personalized tool. Because they can remember things users say, it can get personal and hopefully soon learn soft signals and natural conversation. That way, even if a user doesn't explicitly ask something, a chatbot could potentially know what to recommend based on their patterns or behavior. However, getting smarter over time is a unique challenge in travel, as people's travel preferences can change greatly depending on the purpose of their travel, which means the bot needs to be adaptable to different travel needs.
This is a follow-up to my first post, where I discussed how I built and launched a banking chatbot. I've seen a lot of great discussion lately on sites like Medium and LinkedIn about applying conversational interfaces to various industries and workflows. I hope this article advances that discussion, especially in the field of retail banking.
So think about how to build the whole product with the following in mind: Is the data that's coming out of this product going to be good training data? That, I think, is something that should always be on the table. Some examples of how to do it right would be giving users opportunities to correct errors when they occur, and making sure that it's done in the flow of the product. It should be presented in a way that the user feels like it's going to provide value, because they are helping to fix the product and helping to improve it for their own benefit.
We at Sift Science provide fraud detection for hundreds of customers spanning many industries and use cases. To do this, we have devised a specialized modeling stack that is able to adapt to individual customers while simultaneously delivering a great out-of-box experience for new customers, achieved by mixing the output from a "global" model – trained on our entire network of data – with the output from a customer's individualized model.