Tourism 5.0 isn’t about robots, apps or a hotel lobby full of delivery droids. That’s the shop window. The real story is a new digital architecture being built under travel, and one fight sitting at the centre of it: who owns the layer your travellers’ AI reads from.
Get that wrong and every DMO, hotel and airline will end up supplying data to someone else’s assistant. The sustainability goals will become someone else’s ranking. I’ve been tracking real examples on my Tourism 5.0 guide and tracker, and the pattern is now hard to miss.
Forget the gadgets. Look at the layers.
Most Tourism 5.0 talk is a list of technologies: AI, IoT, digital twins, biometrics, robotics. Put the tracker examples side by side and you see four layers stacking up.
1. The data layer. Destinations are building live models of themselves. The Balearic Islands is creating a digital twin covering beach occupancy, water management and coastal emergencies. In Saudi Arabia, Red Sea Global and stc group have shown a 3D digital twin of Shura Island and Turtle Bay that tracks environment, mobility and marine conditions, and lets teams test the impact of weather or visitor numbers before acting. Kyoto publishes congestion information and live cameras to spread visitors by time and place.
2. The identity layer. At Hamad International Airport, Qatar Airways and SITA have launched Fast Pass, connecting more than 700 touchpoints into one biometric journey. You enrol once and your face becomes your boarding pass. It’s opt-in, and staff are on hand for anyone who prefers standard processing.
3. The agent access layer. This is the one that matters most commercially,. Indonesia’s Ministry of Tourism has launched MaiA, a national AI travel assistant on indonesia.travel that builds itineraries, suggests local food and events, works with maps and handles voice in several languages. At the same time, private companies are opening their booking systems to AI agents through MCP (Model Context Protocol), an open standard that gives an agent one consistent way to read live data and take action. Etraveli opened its flight engine to agents, as I covered in AI agents can book your flights now. On the hotel side, Apaleo says use of its MCP server grew 70% in a quarter.
4. The human and sustainability goal. This is what’s supposed to make it 5.0 rather than 4.0 with better PR. UN Tourism’s World Tourism Day 2026 theme put it bluntly: AI “left unchecked, it can funnel tourists to the same few places and platforms”. Digital twins, crowd data and smart routing are meant to stop exactly that.
Here’s the problem nobody is saying out loud
Layers one, two and four are mostly being built by governments, DMOs and airports. Layer three, the agent access layer, is mostly being built by intermediaries.
That’s the uncomfortable bit. A destination can spend millions on a digital twin that knows a beach is full. But if the traveller’s AI assistant books through an OTA’s connector and reads from an OTA’s listing, that knowledge never reaches the decision. The twin sees the crowd. The agent sends people anyway.
So my view is simple: Tourism 5.0 will be decided by whoever owns the source of truth that agents read from, not by whoever has the smartest sensors. Right now, public bodies are building the brains and private platforms are building the mouth.
MaiA is interesting precisely because it’s a government trying to own the mouth too. Even so, it tells travellers to check prices, schedules, opening hours and current conditions with official sources. That’s honest. It also shows the gap. The assistant can inspire, but the live, bookable facts still sit somewhere else.
We’ve seen this film before. Hotels let OTAs own the shelf, then spent a decade trying to win back direct bookings. Now Booking.com’s AI is answering your guests from your listing. If destinations and suppliers treat agent access as “something tech will sort out later”, the same thing will happen again, at the level of the whole destination.
What DMOs, hotels and airlines should do now
- Treat your data as distribution. Opening hours, prices, accessibility, capacity, rules and seasonality need to be structured, accurate and the same everywhere. Agents repeat what they read, mistakes included, politely and at scale.
- DMOs: publish what the twin knows. If you hold live crowd or capacity data, make a version of it agent-readable. Otherwise your sustainability strategy stays on a dashboard nobody’s AI can see.
- Hotels: check your connectivity. Ask your PMS and booking engine suppliers whether any agent you choose can connect, with proper permissions. If the only route in is through an OTA, that’s your answer.
- Airlines: own the offer, not just the fare. Bundles, change rules and disruption support need to be readable by agents, or the aggregator’s connector will define you as a price.
- Measure whether you’re chosen. Check how AI assistants describe and recommend you today, in ChatGPT, Gemini, Google and national tools like MaiA. Being found matters less than being picked.
None of this needs a robot. It needs boring, accurate, well-structured information and a decision about who gets to serve it. That’s the cheapest Tourism 5.0 investment there is, and the one most likely to be skipped.
The real question
Tourism 5.0 promises technology that serves people and places. I like the idea. But a human-centred, sustainable destination that agents can’t read is just a nice brochure. And a traveller’s AI that reads only from platforms will optimise for the platform.
So who should own the layer travellers’ AI agents read from: national tourism bodies, the suppliers themselves, or the OTAs that got there first? I know where I stand. See the full Tourism 5.0 tracker and tell me where you do.
If you want your brand to be chosen and described accurately in AI answers, see my specialist travel AI and GEO services and wider travel marketing services.
Want more takes like this on AI, search and travel marketing? Subscribe to my LinkedIn newsletter.