AI Readiness statistics for travel brands (October 2026)

Key statistics

• The average hotel website scores 27.6 out of 100 for AI Readiness (Globetrotters.ai, October 2026).
• The average official destination website (tourist office or DMO) scores 30.0 out of 100 for AI Readiness (Globetrotters.ai, October 2026).
• 8.8% of official destination websites and 6.9% of hotel websites are AI-Ready, scoring 45 out of 100 or more (October 2026).
• Only 3.2% of official destination websites declare a valid schema.org type for what they are, against 34.6% of hotel websites (October 2026).
• 16.7% of official destination websites and 32.1% of hotel websites publish an llms.txt file (October 2026).
• Only 0.4% of official destination websites and 0.4% of hotel websites have an llms.txt and an ai.json that pass the AI Readiness quality gate (October 2026).
• 0.1% of official destination websites and 0.0% of hotel websites expose an interface AI agents can call, such as an MCP server (October 2026).
• 32.3% of hotel websites mark up a booking action AI agents can follow, at least partly; 0.1% mark it up fully (October 2026).

AI Readiness of hotels

The average hotel website scores 27.6 out of 100 for AI Readiness, across 2,095 hotel websites in 21 countries (October 2026).

Top 5 hotel websites

Hotel websites by country, against the 27.6 average:

• Ahead (2 points or more above the average): France.
• In the pack: Australia, Austria, Belgium, Canada, Croatia, Denmark, Germany, Ireland, Italy, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom and United States.
• Behind (2 points or more below the average): Czechia, Greece and New Zealand.

Grouped, not ranked: gaps between neighbouring countries are within noise. Full figures in the table below.

AI Readiness of destinations

The average official destination website (tourist office or DMO) scores 30.0 out of 100 for AI Readiness, across 906 websites in 21 countries (October 2026).

Top 5 official destination websites

* Provence Guide is a Globetrotters.ai pilot customer.

Official destination websites by country, against the 30.0 average:

• Ahead (2 points or more above the average): Austria, France and United States.
• In the pack: Australia, Canada, Croatia, Denmark, Ireland, Italy, New Zealand, Norway, Spain, Sweden and United Kingdom.
• Behind (2 points or more below the average): Belgium, Czechia, Germany, Greece, Netherlands, Portugal and Switzerland.

Grouped, not ranked: gaps between neighbouring countries are within noise. Full figures in the table below.

AI Readiness by country

Average AI Readiness score per country and segment, in alphabetical order (October 2026). A country with fewer than 20 measured websites in a segment shows no figure.

What most travel websites miss

Most travel websites do not tell machines what they are: 96.8% of official destination websites and 65.4% of hotel websites declare no valid schema.org type, so AI has to guess (October 2026).

Publishing an llms.txt is not enough: 16.7% of official destination websites and 32.1% of hotel websites publish one, but only 0.4% and 0.4% pass the quality gate together with their ai.json (October 2026).

Almost no travel website can be used by an AI agent yet: 0.1% of official destination websites and 0.0% of hotel websites expose an agent interface. Being read is one step; being able to answer and book is the next (October 2026).

Trend

October 2026 is a new baseline. The scoring rubric and measurement changed since earlier barometers, so these scores are not comparable with previous editions. Month-over-month changes start with the next edition scored on the same rubric.

Methodology

The benchmark audits 3,011 travel websites in 24 countries: official destination websites (tourist offices, DMOs, regional and national tourism boards) and hotel websites. In October 2026, 2,915 were measured and 96 (3.2%) refused our crawler.

Destinations and hotels are scored on their own rubric and reported separately. This page never publishes a blended average of the two.

Each website goes through 45 technical checks grouped into five pillars: machine-readable identity, content for AI, crawlability and access, action readiness, authority and trust signals. Each check returns Pass, Partial or Fail, and the results add up to a score from 0 to 100.

Averages cover 906 official destination and 2,095 hotel websites with a measured score; score bands cover 906 and 2,093. Adoption rates cover 918 official destination and 2,093 hotel websites audited this month.

Scoring version. Destinations: v1.11. Hotels: v1.12.

Limits: the score measures what a website exposes to machines, not what assistants say about the brand (that is AI Visibility). Only the five leaders per segment are named; no other site score is published here.

Definitions

Average score

The mean AI Readiness score (0 to 100) of the measured websites in a segment, weighted by the number of measured websites per country.

Bands

AI-Ready 45 to 100, Developing 25 to 44, Emerging 10 to 24, Absent 0 to 9.

Country figures

Shown only when at least 20 websites are measured in the segment for that country.

llms.txt present

Site serves an llms.txt file at its root that our crawler could read (any quality).

llms.txt quality gate

llms.txt and ai.json together pass our quality gate (combined score 80+). Not a presence measure.

Valid schema.org type

Home page declares the schema.org type that fits the organization (TouristInformationCenter / TouristDestination family for destinations, Hotel / LodgingBusiness family for hotels).

Agent interface

Site exposes an interface an AI agent can call (MCP, A2A or equivalent) and it serves travelers, not site admins.

Booking action

Hotel page marks up a reservation action (schema.org ReserveAction or equivalent) an agent can follow; partial = present but incomplete.

Blocked

Our crawler was refused; the site is left out of presence denominators.

How to cite

Quote any figure with its month and a link to this edition:

Globetrotters.ai, “AI Readiness statistics for travel brands (October 2026)”, https://globetrotters.ai/ai-readiness-statistics/2026-10

Figures are free to reuse with attribution (CC BY 4.0).