AI has shown the potential to revolutionize several industries and the professions they create, like programming, accounting, word processing, and art. We are only beginning to see the tip of the iceberg when it comes to what AI can do, but Google is hard at work fleshing out one of these capabilities — itinerary planning for travel. It is merely a concept, but given how fast generative AI is, well, generating possibilities, you could be using this or something like it ahead of your next vacation or business trip.
Google created this demo app as inspiration for app developers — it wanted to showcase the possibilities of combining its generative AI-powered language model with existing Google APIs. Google believes itinerary planning is the perfect example of a complex, research-intensive task which can be made relatively effortless with AI. So, it has released the open source demo for developers to experiment with and better understand how AI is deployed in this application.
In the demo app, Google first uses the PaLM natural language model API to glean the person’s travel preferences in a conversational tone. The person can define their budget, areas of interest, and preferred activities. The AI can then suggest destinations and poll the Google Maps API for location details and images. After all, images can give you a much better understanding of what to expect from a certain location and its attractions. We had a similar experience using Bard’s newfound ability to display images in search results.
Google’s flowchart for polling the Maps API
However, Google’s travel planner demo goes way beyond. What if you don’t specify a duration for your trip? Five days is the default. Google has also trained the PaLM API to respond in a certain manner, and wait for you to request refinements to the itinerary based on what you’re hoping to do on the trip. The app also has conversational memory, so it already understands context when you circle back to a preference you outlined earlier in the interaction.
Further, Google allows third-party integration, meaning the person planning a trip can book flights and hotels directly in the chat with the AI, after they are satisfied with the tailor made itinerary. Third-party developers can also use the AI to display curated tour packages and destinations based on regularly looked-up keywords like “budget traveler” and “family trip,” saving you the hassle of personalizing the search manually.
Because this whole AI-powered demo is open source, Google is essentially allowing other developers to snoop around its code and possibly even build off the experience in its example app. Since this isn’t walled behind license agreements, it should only be a matter of time before we see AI injecting itself into the best travel planning apps. Google says it is already in talks with partners in Asia, like Agoda, to integrate PaLM into the ticket booking platform’s existing customer flow.
