Journey-rative AI: How can LLMs generate the most perfect itinerary?
Yen Siang Leong
Marketing Strategy & Analytics Lead, Devices & Services and GPay at Google
I have been thinking a lot about how Generative AI will impact my life and recently when i was away for vacation, I found something that excites me and other travelers who enjoy planning for their trip - THE PERFECT ITINERARY! Folks would probably have tried putting together personal itinerary on Google My Maps (as in the Image above) - While that is a really good tool, the effort one has to put in navigating the tools might not be the best part of the trip.
When swapping stories & itineraries with other travelers, it struck me that there is a wide range of itineraries even when everyone is traveling to the same place. Everyone might have followed the same few ‘Suggested Itinerary’ but the things that people actually do vary so much. Depending on how many people you have in your party, who they are, what they enjoy doing, fitness level, the possible permutations increase quickly. Take kids-friendly as an example, every kid is at different development milestones, with different interests - so the broad kids-friendly activities are usually not sufficient. I find myself envying families where the kids are slightly older and could go on longer hikes or not having to stop for nap ??
Lonely Planet/Rough Guides of the world, Suggested Itinerary/Things to do in (insert location) will be able to give a rough idea of the various points of interest, but things like how many days to stay, how to split time between the different things to do, special opening hours, activities requiring additional planning… are not (yet) as easily planned for in advance. This is especially true if one is looking for authentic, local experiences after ticking off the “must-do” checklists.
The best way to maximize your time there is still the old-school way - rocking up to the local tourist information office and talking to locals, going through the brochures and making plans on the fly. While this exudes old-world charm, I wonder if some of this research could be done in advance so that the planning could have been more seamless.?
There have been multiple instances where we wished we could have spent more time than we had planned for in certain places or even simply reordered our itinerary to be able to catch local festivals, markets etc. In hindsight, it is very easy to dismiss it as “We’ve booked too many nights in town A” but the real dilemma is knowing that in advance and the fear of not spending enough time in that place. And if you are someone greedy like me, who wants to see and do as much as possible in 2-3 weeks, you would probably understand this ‘traveler’s remorse’.
There is a growing demand for truly-personalized itinerary planning, especially as ‘revenge travel’ coincides with macroeconomic headwinds. The traditional resources discussed above are good starting points, but insufficient to meet increasingly higher expectations from travelers. Organizations that could provide tailored planning at scale will likely be able to gain share and loyalty with travelers.
领英推荐
Google Search results are illuminating in that the key theme across most industry predictions in 2023 is personalization. These are some examples of the growing macro travel trend, most of which require tailored and bespoke planning, especially if the traveling parties have different interests. These are not necessarily niche trends either, with up to 50% of travelers indicating interest in some of them (Guesses, anyone?)
Independent travelers are still the biggest (80%) & growing segment in the travel rebound. Even within package holidays, the fastest-growing and most lucrative segments are likely within the segments such as wellbeing and indigenous travel. Hence, the quest in finding the perfect itinerary is still relevant for most travelers.
This is where Generative AI might be able to curate tailored itineraries at scale. After that, the other large opportunity is definitely integrating across booking surfaces & things to do and bringing the planning to actual experiencing.
What are the key metrics that are relevant for you when planning for your trip? Please comment so anyone building a product out of this could have the right parameters to design the right user interface.
Thank you for reading!
Master of Integrated Innovation for Products & Services @ CMUiii | Seeking Summer 2025 Internship for Product Management roles
1 年Hello Yen Siang, came across your post while I was doing my research on incorporating generative AI to travel planning apps. This is a great piece written and I would most definitely love to find out more on your point of views, especially your take on some of the existing apps that might be addressing this gap, and how the social component (as highlighted by the comment below) can be further improved as well. Happy to connect and chat more :)
GTM Ops @ Google
1 年Great article & +1 to the endless potential for generative AI in addressing the gap. I always love an "agony" metric. There used to be a site called Hipmunk which optimized flight options based on price, travel time and "agony" (transfers, stops etc). For those willing to spend a bit extra for a more efficent & comfortable route/schedule definitely something to consider. Social sharing aspect to me is also very important. How did my friends with similar travel style to me do it? Can I see on a map which of my friends "endorsed" a restaurant or activity? What did they order? I think there's a gap in the market for a social-type aspect where friends can share past itineraries, with some of the attibutes addressed that you mentioned (kids, group size, budget, pace, adventure level etc).