AI is supposedly taking over every aspect of life. There are custom GPT-assistants for every task imaginable. One of the popular use-cases is building a travel itinerary with AI Trip Planners. I decided to put it to a test for my recent trip to Hong Kong!

I’ve talked about this before in my blog – a well-planned itinerary can make a world of difference to your trip. My definition of well-planned? I’m referring to the hidden gems, the best of local cuisine and truly cultural experiences. It doesn’t mean I exclude the overly touristy places – there is a good mix of both.

With AI trip planners like ChatGPT and Perplexity, it is super easy to create an itinerary. In fact, it is as simple as typing, “Give me a 4 day itinerary for Hong Kong and Macau and include all the hidden gems and authentic experiences”. Or you could just copy-paste the last sentence and replace the destinations. You can sit back and let AI generate detailed day-wise plans for you. While this can be a good skeleton of an itinerary, it falls short. And I mean, VERY SHORT of what I would call an ideal itinerary. All right then, let’s get into the details!

What AI trip planners do well

I like to start with the positives. And there are definitely a couple of those.

1. Speed: The biggest advantage of AI is the speed it offers. Within seconds, I could create the outline of our trip and save hours of manual research. This is particularly useful if you are struggling to decide where to start. Just to give an example, Perplexity told me within seconds the top places to visit in Hong Kong – Victoria Peak, Wong Tai Sin temple, Lantau island and the night markets near Kowloon. This is an excellent place to start.
2. Personalized preferences: I told Perplexity that I wanted to explore the best eateries and I got a bunch of them as recommendations – Australia Dairy Company for breakfast, Sister Wah for lunch and Din Tai Fung for the soup dumplings. You can also get it to list down activities based on your preferences – historical landmarks, family-friendly activities, adventure activities etc.

A good way to support this blog is to subscribe. It’s totally free and all you need to do is drop your email ID and get weekly updates when a new blog is released.

Where AI trip planners struggle

Now that I have a rough list of places, I want to get into some of the finer details. That’s where AI-led research draws a blank or continues to go in loops.

1. Generic recommendations: AI looks at the most written-about places while putting together recommendations. These would invariably be super touristy places. For example, it would definitely add the Eiffel Tower in a Paris itinerary but struggle to tell you the 2 places from where you can get a very unique view of the tower. These are tips which are not part of every website and blog – something that deep research on Google and Tripadvisor can reveal (my hope is that this website is a similar source of inspiration!)
2. Finding hidden gems: Looking for hole-in-the-wall places with mind-blowing local food? You may struggle to find them on those listicles which generally form a part of AI’s base data. These are best found on Google Maps, the smaller niche blogs and through locals.
   For example, we found some really good breakfast places and shopping areas in Hong Kong through local recommendations from friends.

3. Outdated info: This one is a basic issue and a big one. If the data used for training AI is old, you may be served with a bunch of places that are permanently closed. This happened to us for a couple of restaurants in Hong Kong.
4. Optimizing the schedule: While AI is good at providing a list of places, it schedules them based on the data it has seen before. It kept forcing us to make a 5-day itinerary even though the activities could easily fit into 4 days.
5. (TIP: One work-around for this is to ask AI to give you the latitude-longitude of each of the places and then use that information to schedule the plans in a more optimized way. This can overcome some of the “hallucinations”).
6. Cost estimation: I’ve come across many blogs that share “Budget”, “Mid-Range” and “Luxury” stay options in cities and these are all usually super expensive. Possibly to drive more affiliate income. It would be best to stick to Booking.com and Airbnb to find your stay. Even for other activities, Google is the best source for the latest prices. This goes back to point 3 above on outdated info.

Let’s talk solutions

Now that we’re done with the pros and cons, let’s move into solution mode. Nothing beats detailed research to complete your planning. Here are some resources that I rely on:

• Tripadvisor threads – these are the absolute best to find opinions on itineraries as well as hidden gems. People are actually very helpful on these forums and you can even get answers to queries about visas and the real experiences of others.
• Reddit threads – similar to Tripadvisor but with more Gen-Z participation. Just be careful that you don’t go down random rabbit-holes here!
• Advanced Google search – using the word “blog” in your search can help weed out most of those paid listicles. If it doesn’t, page 2 and 3 of search results definitely takes you to those “hidden gems” I spoke about earlier.
• Google reviews – a final checkpoint to make sure you pick the right places. This can help become a tie-breaker if there are too many places and too little time.

I don’t lean too much on reels and shorts as there may be an economic incentive for the person posting those (sponsorship, engagement etc). There are also too many instances of them being “manufactured”. 

Time to summarize!

What a cliched ending no? Here too, a balance of AI trip planners and human research is the most optimal solution!

• Start with AI and create your skeleton itinerary
• Dive deeper with your own research on Tripadvisor, Google and Reddit
• Validate the plan with Google reviews
• Engage with locals to make the plan even better
• That’s it! 

I used to be totally biased towards manual planning as I derive a lot of joy from the deep research. And the best part is that it does not take away from the fun of actually exploring the place. How do I do it?

Avoid pictures and videos! Like I said above, I avoid reels and videos as much as possible. Even on Google, I don’t go looking at a lot of photos of the place. I only read as much as I can and form a mental picture. So, the final reveal is always exciting!

But I’m also biased to learning new stuff and optimizing (you’ll know this if you’ve read my blogs on new year resolutions and overcoming my resistance to writing)! So, AI has become a constant starting point for my itineraries in the last 1.5 years. The Georgia trip, Armenia trip and Hong Kong trip have all had AI trip planners playing their role.

What about you? Do you prefer the ease of AI tools or the depth of personal research? Or a bit of both like me? Share your thoughts in the comments! Also, if you’re looking for curated itineraries of the following countries, do check out these blogs:

• Bhutan
• Cambodia
• Jordan
• Kazakhstan
• Sri Lanka
• Turkey
• Dubai
• Uzbekistan
• Vietnam
• Armenia
• Austria
• Croatia
• Prague
• Georgia
• Budapest
• Liechtenstein
• Monaco
• Vatican City
• Egypt
• Seychelles