BALANCED TOURISM

Artificial intelligence has already taken over the question a destination marketing organization used to answer for itself: what should a visitor do here? The organizations that treat that as a promotion problem are going to lose the channel. The ones that treat it as a data problem are going to run it.

By Dr. Jens Thraenhart | CEO, Chameleon Strategies (UN Tourism Affiliate Member)

For most of my career, a destination marketing organization’s core job was legible in its name: market the destination. Build the campaign, place the media, tell the story, count the arrivals. I ran that playbook at Destination Canada, at Fairmont, at the Mekong Tourism Coordinating Office, and as CEO of Barbados Tourism Marketing Inc. It worked because the DMO sat, structurally, between the traveler and the information about the destination. If you wanted to know what to do in Barbados, you eventually ran into something BTMI had produced, funded, or influenced.

That structural position is gone, and it is not coming back. A growing share of travel research now happens inside a conversation with an AI system rather than on a destination’s own website, and the DMO is no longer reliably in that conversation. Skift’s own 2026 data and AI reporting found that only 6% of hotels currently appear when travelers query AI search tools, a number that should alarm anyone whose job depends on being found (Skift, 2026). On the demand side, PwC’s May 2026 hospitality outlook found 44% of United States travelers already use AI tools to compare prices, and roughly a third use AI agents or bots to actually book (PwC, 2026). Destinations are being described, ranked, and recommended by systems that most DMOs have no structured relationship with at all.

Disintermediation Is a Data Problem, Not a Marketing Problem

The instinct inside most DMOs, mine included at various points, is to respond to a discoverability problem with a campaign. That instinct is the wrong reflex here. An AI system does not discover your destination through a beautifully produced hero video. It discovers your destination through structured, machine-readable information: what is open, what it costs, what it is near, what people who actually went there said about it. A destination that has not organized that information for machine consumption is not losing a marketing battle. It is simply absent from the dataset the machine is drawing from, and a third party fills the gap with whatever it can find, accurately or not.

I watched a version of this problem in an earlier form in Barbados. The job was never really “promote the beaches.” Global sentiment data made it obvious the beaches were already working; Barbados ranked in the top 10% of destinations worldwide for beach sentiment, and outranked a direct regional competitor for five straight months of 2021 by more than 20 points (Destination Think, 2022a, 2022b). The actual job was managing what the data showed underneath that headline number, and deciding where to invest attention that the beach score alone would never reveal. That is a data and governance function wearing a marketing department’s job title. AI has just made the gap between the two impossible to ignore.

What Continuous Sentiment Already Proves

Destination management has relied on the visitor satisfaction survey for decades: a periodic, retrospective, sampled instrument that tells you, months later, how people felt. That instrument is being overtaken by continuous, machine-scored sentiment drawn directly from what travelers are already saying, at a scale and a cadence no survey can match.

The Hong Kong Polytechnic University’s Research Centre for Digital Transformation of Tourism, under Professor Haiyan Song, built exactly this. Its large-language-model Hong Kong Tourist Satisfaction Index processes more than 1.25 million online reviews across nearly 13,700 service providers, scoring satisfaction by sector, district, and trip type every month rather than once a year (The Hong Kong Polytechnic University, 2025; Li et al., 2023). That is not an incremental improvement on the survey. It is a different category of instrument, and it exists today, not as a concept paper.

I used an earlier version of the same idea in Barbados with Destination Think’s Tourism Sentiment Index, which scores destinations from public online conversation rather than solicited responses. The value was never the single number. In the first quarter of 2022, our score was 28, essentially flat against the prior quarter, and a headline read of that number alone would have told a manager to keep doing what they were doing. The topic-level detail said something different: beaches were driving a large share of the conversation at a strong sub-score, but air travel’s sub-score sat well below it, flagging a friction point in the visitor experience that the composite number buried entirely (Destination Think, 2022a, 2022b). A satisfaction survey run twice a year would have caught that eventually. The continuous signal caught it in real time, while there was still a season left to act on it.

A destination that has not organized its information for machine consumption is not losing a marketing battle. It is simply absent from the dataset the machine is drawing from.

The Governance Gap Is the Real Barrier, Not the Technology

Here is the finding that should reorder how boards and city councils think about this. Technology cost is not what is stopping destination organizations from doing this well. A small-city DMO in New Westminster, Canada runs a 24/7 AI visitor assistant, saving roughly 10 staff hours a month, on a technology stack costing USD 210 a month, run by a 2.5 person team (Tourism AI Network, n.d.). Venice’s much larger Smart Control Room cost around EUR 6 million (Lepschy, 2025). Both are affordable relative to the value they generate. Neither budget is the constraint.

The constraint is governance capacity: the staff, the mandate, and the internal authority to decide what data an AI system may use, how visitor and resident privacy is protected, and who owns the relationship with the vendor once the contract is signed. City Destinations Alliance’s 2025 research on European city DMOs found only 29% now retain full-time sustainability staff, down from 42% just two years earlier, even though 94% of respondents called the work personally urgent (City Destinations Alliance, 2025). Layer AI governance on top of a sustainability function that is already shrinking, and you get destinations adopting tools faster than they can supervise them, which is close to the worst version of this transition.

From Marketing Department to Data Steward

None of this argues that DMOs stop marketing. It argues that marketing becomes one function inside a broader mandate: managing the destination’s data, its discoverability, and its relationship with the AI systems that now sit between it and a meaningful share of demand. Three shifts follow directly from the evidence above, and none of them require a large budget to start.

• Publish for machines, not just people. Structure official content, hours, pricing, accessibility information, and event data in an open, machine-readable format, the same discipline behind Flanders’ linked open data model for tourism. A destination that will not describe itself accurately to an AI system is choosing to let someone else do it with worse information.
• Replace the annual survey with a continuous signal. Adopt AI-scored sentiment tracking, whether an academic model like PolyU’s or a commercial one like Destination Think’s, as a standing management input rather than a one-time report. The value is not the score. It is the topic-level detail that shows up months before a survey would.
• Fund governance before the next procurement, not after. Settle data rights, privacy design, retention rules, and vendor continuity as a written protocol before signing the next AI tool contract, not while troubleshooting the first one. This is a staffing and mandate decision more than a technology decision, and it is the one destinations are currently underfunding.

Stewardship Is the Job Now

I do not think this is a loss for destination organizations. It is a more interesting job than the one I started in. Promotion is something you can outsource to an agency or, increasingly, hand to an algorithm that has read every review ever written about your destination. Deciding what data an AI system should be trusted with, what a continuous sentiment signal is actually telling you, and where governance capacity needs to sit inside a small organization: that requires judgment a machine does not have, and it is squarely a management function, not a communications one.

The destinations that make this shift early will not just be more visible in AI search. They will be the ones with an accurate, current, well-governed account of themselves circulating in every system that now mediates a traveler’s decision. The destinations that wait will be described anyway, just not by anyone who works for them.

About the Author

Dr. Jens Thraenhart

Dr. Thraenhart is CEO of Chameleon Strategies (UN Tourism Affiliate Member), Founder of Saudi Outbound, Co-founder of High-Yield Tourism, Author of the Passion-Tourism Economy, and an Advisor to the Saudi Tourism Authority. His prior roles include CEO of Barbados Tourism Marketing Inc., Executive Director of the Mekong Tourism Coordinating Office, Executive Director of Marketing Strategy at Destination Canada, and Executive Director of Digital Strategy at Fairmont Hotels and Resorts. He co-founded Dragon Trail China, one of the earliest firms focused on digital marketing for Chinese outbound tourism.

Sources

Skift. (2026, July 1). Skift Data + AI Summit 2026: 10 insights from travel’s AI frontlines. Skift. https://skift.com/insights/skift-data-ai-summit-2026-10-insights-from-travels-ai-frontlines/

PwC. (2026, May 28). US hospitality directions, May 2026. PricewaterhouseCoopers. https://www.pwc.com/us/en/industries/consumer-markets/hospitality-leisure/us-hospitality-directions.html

The Hong Kong Polytechnic University. (2025, April 25). PolyU unveils large language model-based tourist satisfaction index, providing comprehensive analysis to enhance Hong Kong tourism service quality [Press release]. https://www.polyu.edu.hk/media/media-releases/2025/0425_polyu-unveils-large-language-model-based-tourist-satisfaction-index/

Li, H., Gao, H., & Song, H. (2023). Tourism forecasting with granular sentiment analysis. Annals of Tourism Research, 103, Article 103667. https://doi.org/10.1016/j.annals.2023.103667

City Destinations Alliance. (2025). The score is not the story: A critical reflection on sustainability indexes and certifications in city tourism management [VivaCITY whitepaper]. City Destinations Alliance. https://citydestinationsalliance.eu/initiatives/vivacity

Tourism AI Network. (n.d.). Case study: Tourism New Westminster’s AI visitor assistant.

Lepschy, F. (2025). The Venice Smart Control Room: Governance and cost of a municipal monitoring platform.

Destination Think. (2022a, March). Barbados & TSI presentation [Client presentation prepared for Barbados Tourism Marketing Inc.; not publicly available].

Destination Think. (2022b, March). Barbados tourism sentiment snapshot: Jan 01–Mar 31, 2022 [Client report prepared for Barbados Tourism Marketing Inc.; not publicly available].

Destination Think. (n.d.). Tourism Sentiment Index: The data you need to lead your destination. https://destinationthink.com/tsi/tsi-tourism-sentiment-index-data-destination-marketing/

Estimates cited without a named publication represent figures drawn from the author’s own consulting engagements and are qualified as such in the text.