How to ask AI to create the perfect travel itinerary

SKIP TO CONTENT

Tiếng Việt

How to ask AI to create the perfect travel itinerary

Dr Daisy Kanagasapapathy, Interim Associate Program Manager of the Tourism and Hospitality Management program at RMIT Vietnam, explains why AI remains a work in progress and how better prompts can improve travel planning.

Why AI travel recommendations often feel generic

The primary reason AI tools, such as ChatGPT, Google Gemini, or AI-based travel planners, tend to recommend popular destinations is that they are trained on widely available and frequently mentioned content. When millions of travel blogs, social media posts, and tourism websites all talk about the same places, such as “Eiffel Tower” or “Ha Long Bay”, those destinations dominate the training data. This leads AI to associate popularity with relevance, even when users want something off the beaten path.

Younger travellers often seek authentic, hyper-local, or niche experiences that reflect their personal identity or social values, such as vegan-friendly cafes in Ho Chi Minh City, zero-waste homestays in Da Lat, or underground music venues in Seoul. However, these types of recommendations are more complex to surface unless they are discussed in depth online, reviewed frequently, or tagged consistently.

In short, AI reflects the mainstream, unless prompted specifically. That explains why Gen Z travellers often use AI for rough drafts, then dig deeper using TikTok, Threads, or Facebook groups to personalise their plans.

Dr Daisy Kanagasapapathy, Interim Associate Program Manager, Tourism and Hospitality Management, RMIT University Vietnam

Why AI sometimes gets it wrong: Outdated or inaccurate suggestions

From a technical perspective, there are two common reasons AI provides outdated or incorrect travel suggestions.

Firstly, cut-off dates in training data are still a common limitation. Many generative AI models are trained on data that only reaches a certain point. For example, ChatGPT’s free version was previously trained only up to late 2023. This means AI won’t “know” about recent changes, like a restaurant closing last month, a new museum exhibit, or an updated ticket policy.

Secondly, AI lacks real-time integration. AI tools are not search engines by default. Unless connected to live web sources, such as via plugins, APIs, or real-time web access, they cannot pull current prices or availability. So, if you ask an AI, “What’s the current entrance fee for Bana Hills?”, you might get an average from two years ago, not today’s rate, which may have changed due to seasonality or promotions.

Additionally, user-generated travel content can be biased or unverified. If AI has read a dozen outdated blog posts about a place that closed during the pandemic, it might still recommend it, unaware that it’s no longer operating.

While AI isn’t perfect, it can still be a powerful tool when used wisely to create travel itineraries. (Image: Freepik)

How to use AI effectively for realistic, personalised itineraries

While AI isn’t perfect, it can still be a powerful tool if it is used wisely. Here are a few practical tips for Gen Z travellers and beyond:

Start with AI but don’t stop there. Use AI to create a basic framework: “3-day food itinerary in Osaka” or “Eco-tourism route in the Mekong Delta.” Then cross-check on platforms like Google Maps, TripAdvisor, and TikTok for recent reviews and visuals.

Be specific in your prompts. Instead of asking “What should I do in Ho Chi Minh City?”, try: “Suggest a 2-day itinerary in Ho Chi Minh City with street food, local art galleries, and public transport options.” The more details you provide (budget, travel style, dietary needs), the better the output will be.

Ask AI to simulate or compare options. For example, “Compare Mui Ne and Quy Nhon for a quiet beach trip with local seafood,” or “What are some lesser-known spots in Ninh Binh popular among university students?”

Verify time-sensitive details manually. Always double-check opening hours, ticket prices, and seasonal availability through official websites or Google listings. Treat AI as a starting point, not a replacement for a booking agent.

Look for AI tools that integrate live data. Some travel platforms now offer AI-powered suggestions combined with real-time updates, including Expedia, Kayak, or Klook’s innovative planning tools. These are more reliable for pricing and availability.

Travel planning as a hybrid journey

For Gen Z, planning a trip is almost as exciting as the journey itself. AI offers speed, convenience, and a helpful starting point, but it can’t (yet) fully replace human intuition, local knowledge, or lived experience. That’s why young travellers still rely on community-generated platforms and digital storytelling to add that final personal touch.

In a way, AI is like a well-read but distant friend: it knows a lot, but it doesn’t know you. The trick is learning how to ask and when to explore beyond the algorithm.

Story: Dr Daisy Kanagasapapathy

Media Release

07 August 2025

Share

Engage

Contact RMIT News

Find an expert

RMIT Vietnam facts & figures

Related news

Person's hands holding two containers of food

Trust is the most valuable ingredient in “ghost kitchens”

Business and management

06 Aug 2026

New RMIT University research reveals what drives trust in delivery-only food businesses, offering a roadmap for operators in the digital food economy.