The short version
More than half of US leisure travellers used AI to plan at least one trip in the past twelve months — up from 43% in late 2025 and 24% in 2024. Phocuswright calls it the fastest behavioural shift travel has seen in a decade.
For tour and activity suppliers, that creates a simple problem. A traveller asks ChatGPT for the best food tour in Seville. An answer comes back. If your product is not in it, you were never in the running — and you will never see the impression in your analytics.
This is a distribution shift, not a marketing trend. Here is what the data says, and what suppliers should do about it.
How fast is AI travel planning actually growing?
Fast, and from every direction at once.
| Metric | Figure | Source |
|---|---|---|
| US travellers who used AI for at least one trip (past 12 months) | 56%, up from 43% in late 2025 and 24% in 2024 | Phocuswright, March 2026 |
| Year-on-year growth in generative AI use for trip planning, globally | +64% | Amadeus Travel Trends 2026 |
| Adoption among travellers aged 18–35 | 71%, up from 52% in 2024 | Phocuswright via ALPN |
| Travellers who say AI improved their experience | 84% of those who used it | McKinsey |
Two things stand out. The first is the slope: 24% to 56% in roughly eighteen months is unusual for any technology category in travel. The second is who is moving. Millennials sit at 74% and Gen Z at 72% — the cohorts that over-index on experiences rather than package holidays.
Meanwhile the underlying market keeps growing. Tours and activities reservations were valued at $185.5bn in 2025 and are forecast to reach $199bn in 2026, growing at 7.8% a year. The category is expanding and the discovery layer in front of it is being rebuilt at the same time.
Why does this hurt tours and activities more than hotels or flights?
Because tours are a discovery-led purchase.
Nobody asks an AI assistant which airline flies to Lisbon — they already know, and they go straight to a booking site. But "what should I do with three days in Lisbon?" is exactly the open-ended, preference-heavy question that travellers now hand to an assistant. McKinsey found general research is the top AI travel use case at 54%, with travel inspiration and local recommendations close behind.
That is the top of the tours and activities funnel, and it is moving inside a chat window.
The second reason is fragmentation. Hotels and airlines are represented by a handful of large, well-structured inventory sources. Tours and activities are tens of thousands of small suppliers whose product data lives in reservation systems that were never designed to be read by a machine. When an assistant cannot parse your availability, pricing, or cancellation terms, it does the safe thing: it recommends someone it can parse.
The trust gap is the real opportunity
Adoption is not the same as conversion, and the gap between them is where the commercial opening sits.
Travellers use AI for recommendations, itineraries and ideas, then revert to old habits at checkout — only 13% trust AI for the actual booking. Around 25% report having received outdated or incorrect information from an AI tool while planning, and only 46% are willing to trust AI with travel decisions.
Read that from the supplier side rather than the traveller side. The reason assistants hedge is that most of what they can see about a tour is stale: a price scraped from a listing page, no live availability, no confirmation that the operator still runs the departure. The models are being cautious because the data underneath them is unreliable.
Suppliers who expose accurate, live, structured data are not just easier to recommend. They are safer to recommend. As assistants move from suggesting to transacting, that difference decides who gets surfaced.
What AI visibility actually requires
Being "on the internet" is not the same as being readable by an AI system. In practice, visibility comes down to four things.
Structured, machine-readable product data. Schema markup and clean structured data so an AI crawler can identify what the experience is, where it runs, how long it lasts, and what it costs — without guessing from prose.
Live availability and pricing. A recommendation an assistant cannot verify is a recommendation it will soften or skip. Static listing pages age badly and are treated accordingly.
Consistency across sources. When your duration, meeting point and price differ across five OTA listings and your own site, an assistant has no canonical answer to give. Conflicting data is worse than thin data.
Demonstrable reliability. Response speed, booking success rate and reputation sentiment are all signals of whether a supplier will actually deliver. At Tixxly we combine these into a Trust Score — a four-factor metric covering data quality, response speed, booking success rate and reputation sentiment — because the systems doing the recommending increasingly need a reason to pick one operator over another.
What suppliers should do now
- Audit what AI assistants currently say about you. Ask ChatGPT, Gemini, Perplexity and Claude for the best experiences in your city and category. Note whether you appear, whether the details are right, and who is being recommended instead. This is your baseline.
- Fix the data before the marketing. Correcting inconsistent durations, prices and meeting points across your own site and your distribution partners is unglamorous and it is the highest-leverage work available.
- Make availability machine-readable. If an assistant cannot check whether Thursday at 10am is bookable, it cannot confidently send anyone your way.
- Treat AI referrals as a measurable channel. Segment traffic arriving from AI assistants in your analytics. If you are not measuring it, you cannot tell whether any of this is working.
- Connect once rather than integrating repeatedly. The number of AI surfaces is growing. Maintaining a separate integration for each one does not scale for a supplier with a small team.
Where Tixxly fits
Tixxly is AI infrastructure for travel commerce. We connect tour and activity suppliers to AI ecosystems — ChatGPT, Gemini, Perplexity, Claude and others — so their products can be found, understood and booked when travellers ask an assistant for recommendations.
We are not an OTA. We take no commission on bookings, so suppliers keep 100% of booking revenue less card processing fees. Connect once, reach everywhere.
The discovery layer for experiences is being rebuilt right now. The suppliers who make their data legible to machines during this window will be the defaults that assistants reach for once the shift is finished.