This was created with no code leveraging Rainbird's new workflow and connectivity functionality and its ability to dynamically design interactions. Looking at the interaction, it's easy to imagine there is a decision tree behind this Digital Person - but there isn't. Rainbird is dynamically using "what it knows" in its model (in this example, what makes a good holiday) and dynamically creating non-linear questions to gather the data that it needs to make a strong inference. If it had some of the data already, it would have created a different more efficient interaction.
If you would like to see "under the hood", get in touch. https://vimeo.com/610536282/12b09de4f6