Untangling Tourism Tech — Podcast

Episode 20: What AI Search Means for DMOs

3 July 2026

Listen on your favourite platform

Destination marketing organisations are watching their website traffic fall by an average of 30%, and Google Analytics can’t tell them why. The content is still working. Travellers are still finding it. They just aren’t clicking through anymore, because an AI tool has already given them the answer.

In episode 20 of Untangling Tourism Tech, Fabienne Wintle and Liz Ward unpack what AI search means for DMOs: why the old measures of traffic no longer tell the whole story, what server logs show that Google Analytics doesn’t, and the practical framework in Tourism Tribe’s DMO AI Playbook for rebuilding value in this new search environment.

The zero-click search problem hitting DMOs

When someone asks ChatGPT or Gemini what to do in a region for a weekend, the answer comes straight back inside the chat. There’s often no link to click, no list of ten options like a Google search would return. The traveller gets served an answer built from a DMO’s content, and the DMO never sees it in their analytics.

That’s the “zero-click surge,” and it’s flipping how DMOs have always proven their worth. Visits, page views, and time on site have long been the numbers DMOs report to boards and funding bodies. When a visitor never lands on the page, none of that shows up, even though the DMO’s content did the work.

Reading server logs when Google Analytics goes quiet

To see what AI tools are actually doing with a DMO’s content, the answer isn’t Google Analytics. It’s the server logs, one level upstream of where analytics tools sit.

Two types of bots show up there. Indexing bots, the AI equivalent of Googlebot, visit regularly to read and store a site’s content so it can be referenced later. Retrieval bots behave differently: they’re dispatched in the moment, during a live query, to fetch a page’s schema and metadata and serve an answer back to whoever asked. Both leave a trail on the server. Neither shows up as a website visit.

A tool like Cloudflare, sitting between a server and the internet, makes this visible with graphs and charts of bot activity over time. It’s also where a DMO can filter out bad bots without accidentally blocking the ones that matter, since blocking ChatGPT’s bot means disappearing from ChatGPT’s answers altogether.

The Compass Model and Tourism Tribe’s DMO AI Playbook

Tourism Tribe released the DMO AI Playbook a few weeks before this episode, built around a practical framework called the Compass Model. It gives DMOs a plan for reinventing their role and protecting their relevance as search moves toward AI.

The first step in the model has nothing to do with technology. It’s leadership, getting a board and decision-makers to genuinely understand what’s changing. One of the simplest ways to do that is to hand a board member a phone and let them have a live conversation with ChatGPT about their own destination.

Crawlable versus queryable content

Crawlable is what sits above the waterline: the content published on a website, and off it, that AI engines can find, read, and cite. That’s the part most DMOs already think about. Queryable is what sits below the waterline, and it’s where the real opportunity is.

Queryable content is the knowledge a DMO already holds but has never digitised: itineraries built for past journalist visits, workshop notes, member relationships, years of local expertise sitting in people’s heads rather than in a searchable system. Turn that into an organised knowledge base, and a request that used to take hours, like preparing an itinerary for a visiting journalist, can be drafted automatically by AI pulling from what’s already been done before.

The starting point is simple: digitise everything. Meetings get transcribed. Notes get stored somewhere searchable. Over time, that becomes the foundation an AI-enabled organisation can build on.

The travel purchase cycle is collapsing into one session

ChatGPT trialled in-chat shopping, then pulled back from travel specifically, leaving fulfilment to the OTA apps built for it. Too many variables, too much complexity. Google went the opposite direction and built the purchase experience directly inside its own AI tools, backed by existing relationships with OTAs and a head start on knowing the consumer through products like Gmail.

The practical effect: a traveller can research, decide, and book inside a single AI conversation, without ever landing on a DMO or operator website.

Understanding AI agents

Talking to ChatGPT about a holiday is closer to briefing a personal assistant than searching a website. The assistant knows your preferences, does the research, and comes back with a decision already made. That’s what an AI agent is: something acting on your behalf, with your authority, to get a task done.

Applied to booking, it means the shopping cart is already filled before a person reads a single review. Adoption of fully autonomous booking is still low today, but the investment behind it isn’t small. A large share of OTAs are reportedly planning to deploy agentic booking at scale, so the shift from low adoption to normal practice is likely to move fast.

Why DMOs need to move now

DMOs used to be the middlemen between a traveller and a destination’s businesses. AI has made that role harder to justify twofold: it’s easier than ever to get information directly, and increasingly, no one visits a site at all because an agent, especially one built by Google, already knows enough to close the deal.

What doesn’t disappear is the accuracy of what an LLM says about a region. A DMO’s job shifts from displaying content on its own website to making sure the destination is represented accurately wherever an AI agent is drawing from.

Case studies: Toronto and Brand USA

The DMO AI Playbook includes case studies on Toronto and Brand USA, both further along on AI adoption and building the internal skills to sustain it. Neither example requires a Brand USA-sized budget to apply. Any DMO that values its local knowledge and its member relationships already holds the core asset. What’s missing for most is digitising it, laying the foundations, then adding AI tools in incremental steps once that groundwork is done.

The DMO AI Playbook, including the Compass Model, is available at tourismtribe.com/compass-ai-playbook-dmo. Tourism Tribe is also happy to run a quick debrief on the playbook for any DMO working out where to start.

What is zero-click search and why does it matter for DMOs?

Zero-click search happens when someone asks an AI tool like ChatGPT or Gemini a question and gets a full answer with no link back to the website that supplied the information. DMOs are reporting an average 30% drop in website traffic because of it, even though their content is still being used to answer travellers’ questions. The visits just don’t show up in Google Analytics.

Why doesn’t AI search traffic show up in Google Analytics?

AI tools use two kinds of bots to serve their answers: indexing bots that regularly crawl and store website content, and retrieval bots that fetch a page in real time when a live query needs an answer. Neither type brings a human visitor who clicks through, so none of it registers in Google Analytics. The activity is visible in server logs instead.

What’s the difference between crawlable and queryable content?

Crawlable content is what’s published on a website, and off it in places like LinkedIn posts, that AI engines can find and reference. Queryable content is the knowledge that lives inside an organisation, such as staff expertise, past itineraries, and workshop notes, that has been digitised and organised so AI tools can search and use it directly.

What is the Compass Model in Tourism Tribe’s DMO AI Playbook?

The Compass Model is the practical framework in Tourism Tribe’s DMO AI Playbook for reinventing a destination marketing organisation’s role in an AI-driven search environment. It starts with leadership, getting decision-makers to understand what’s changing, then moves through digitising internal knowledge and building AI-enabled processes in incremental steps.

How is the travel purchase cycle changing?

ChatGPT tested in-chat shopping but stepped back from handling travel bookings directly, leaving that to OTA apps. Google has moved the other way, building a purchase experience directly inside its own AI tools. Combined with rising investment in agentic booking across the OTA sector, the traditional multi-step travel purchase cycle is compressing into a single AI-assisted session.

What should a DMO do first to prepare for AI-driven search?

Start by getting the destination’s content into a form AI engines can crawl and cite, and get leadership genuinely engaged, for example by showing board members a live chat conversation with ChatGPT about their region. From there, focus on digitising the internal knowledge already held by staff and volunteers, since that local expertise is the core asset AI can’t replicate.

Read the full transcript of Episode 20

Untangling Tourism Tech Episode 20, with Fabienne Wintle and Liz Ward. Each line shows who is speaking and when. Times match the video. We removed filler words and fixed mis-heard names.

[00:02] Liz Ward: Hi, Fab.

[00:04] Fabienne Wintle: Hi, Liz.

[00:06] Liz Ward: Untangling Tourism Tech podcast, Episode 20. That’s awesome. And despite being super busy trying to keep up with everything that’s going on because of AI, we’re back and talking about it and sharing. And today, firstly, welcome to all of the Tourism Tribe members and other listeners. We’re really pleased today to have a discussion about what is going on in the travel purchase cycle and the impacts on destination marketing organisations.

[00:42] We have spent a lot of time, a lot of our focus over the last few months has been working with organisations where they’ve asked for guidance about what is going on, what should they be looking at, what is their roadmap going to look like going forward with a big underscored objective as to how do they maintain their value in that ecosystem that they are supporting, all those wonderful regional areas in tourism where the destination marketing organisation plays such an important role in supporting the sustainability and the collaborative relationships and the marketing for their region. So we’re going to tackle a few questions. Now, Fab, because you’ve been doing a lot of these consults and are really close to the research and the answers, I’m going to be playing interviewer a lot of the time here.

[01:41] But I know we’ll end up back and forth a bit. But how about we start by having a chat about what is actually going on. And this is straight from the industry and from the analytics that we can see in that we know that the DMOs are experiencing a tangible impact on visitation to their websites. What are you hearing and seeing?

[02:09] Fabienne Wintle: So I’m hearing quite a… I mean, in the world of AI and research, it’s very obvious, you know, zero-click search, invisible search. But what does that mean? When I talk to staff at tourism organisation, I realise that, you know, there’s a great divide between what’s happening in AI and what does it actually mean practically for tourism organisation. And the first one, the understanding that when someone goes on an LLM or ChatGPT, you know, Gemini, and searches for something, there is often no link that takes them back to the website that holds the information discussed there. So if someone does a search about what shall I do in Melbourne with three kids over the weekend, right, before in Google, you’d see 10 options and you’ll be able to click on it and land on these businesses’ websites. That doesn’t happen now.

[03:06] That’s what we mean with a zero-click search. There is no ability to click. So it makes it really hard for tourism organisations to actually explain their value because you will not see that in Google Analytics because they didn’t actually come to the DMO’s website. So how do we measure the value, the contribution of the work that the tourism organisation does because the model has been totally flipped? You know, they don’t visit the pages. They don’t visit the listings. They don’t land on Google Analytics. And how do you prove your worth as a tourism organisation?

[03:45] Liz Ward: And on average, we’re hearing, we’re seeing that there is a 30% drop in traffic to DMO websites on average, some more, some less. But it’s a serious impact on the DMOs in terms of how they measure value and report to funding organisations, to their board, to the industry, which is what is getting people so concerned about how they reinvent their role. So this zero-click search, invisible search, I love it. New language that we’re now working with when it comes to digital marketing. If they know that they’re down on website visits, what, and that’s just looking at Google Analytics, how else can they measure their value if they are being queried or are being searched and discovered in the LLMs?

[04:36] Fabienne Wintle: Well, there’s another new world again. This is obviously AI. In order to be able to serve your content, they have to come and fetch it somehow. And all the LLMs have got bots, like we’re familiar with Googlebot, you know, in the old days, well, it still exists. That’s what would go to your website, read it, and then, you know, decides whether or not it would be listed on the search engine. Well, AI have got bots as well. But these bots, they visit your website server, they don’t bring traffic to your Google Analytics, because people don’t follow links. So you have to go up river, upstream once, and actually look at the server logs. So a bit harder, you might have to wrangle some data, and then see how often and which bots come to read your content regularly, so index your content.

[05:27] So ChatGPT, Gemini, they come and index. And then the other type of bots are the ones who come and retrieve your content in a live query on an LLM. So you know how LLMs, they index your content regularly, but then if they can’t find what you’re asking for, they will send a bot quickly to crawl your website, read your schema, right, this is what’s important, because it can’t read the whole website, it reads the meta information, to then serve it back in the query. So these two different types of bots, they will all be represented on your server logs.

[06:03] And if you use a tool like Cloudflare, if you work with us, you know that we set everyone on Cloudflare, that’s an in between, between your server, you’ll be able to then see inside Cloudflare some nice graphics and charts, and hopefully your growth in visits by these different types of bots. That’s also where you can block the bad bots, but make sure you don’t block all of them, right, because you certainly don’t want to block like ChatGPT, otherwise you’re not going to appear there. And on many of our sites and sites that clients will work with, we’ve seen a huge increase in the last few months of, you know, these retrieval bots that come on the spot.

[06:46] So server logs instead of just Google Analytics, don’t despair because you’ll see a drop in visits to your Google Analytics, but it doesn’t mean that your content is not useful, but to prove that you’ve got to go and look at the server logs.

[07:03] Liz Ward: Absolutely. And I would have thought that it would apply to all of the industry and businesses, that those people who are visiting your website because of the AI environment that they’re searching in now, they’re warmer leads anyway. So that’s a benefit, that’s a trade off, less visits to the website, but those that have come have already been exposed to an AI summary or similar to get there, which is great. Now, not that long ago, about six weeks ago, we released the DMO playbook for AI. And that came about because we knew that we needed to bring this information together as well as a roadmap. What would a plan look like to be able to reinvent the value and secure the ongoing relevance and sustainability of a DMO? So we put the DMO AI playbook together.

[08:06] We’ve had some really lovely sentiments of gratitude and great feedback on that, which is terrific. And one of the concepts that really led in the model that we articulated in the playbook, the Compass model, so we put it basically a model that people could follow to create a plan. And one of those first concepts is about getting this valuable content that all the DMOs have out of the heads of the amazing staff and teams, VIC volunteers, and making it queryable. There’s also this talking new language again in this AI environment that we’re in, queryable versus crawlable. So Fab, you’ve been talking about websites being crawlable and having schema that can be very quickly searched on and retrieve information. Tell us about this concept of crawlable versus queryable.

[09:11] Fabienne Wintle: That might be a Fabi concept, but it works really well.

[09:14] Liz Ward: I think we should keep it as language for everybody.

[09:18] Fabienne Wintle: Okay, so imagine an iceberg. What’s above the waterline is being AI crawlable. And that’s what we were talking about before, Liz, making sure that your content that is on your website and online, and let’s say you write LinkedIn posts, all that jazz, can be crawled by AI engines in order to retrieve your content on an LLM. Really important, being visible to the AI engine. But what’s underneath the surface and what I’ve been doing with you and the team for so long now is making sure that a whole organisation can benefit from all the novelties and the efficiencies of AI. So that’s what I call under the waterline.

[10:03] And it means that everything that we do, our knowledge, all our workshops, our podcast, and similar for DMO, like think about the content they create, but also the itinerary for media, the proposals, the education that they do with their members. All that takes time, right? Let’s say you receive an itinerary request from journalists. There’s no need to start from scratch again. You can use AI to crawl your inbox, index all the past itineraries that you’ve had in the past. Past itineraries in the past, yes indeed.

[10:39] Put it in a brain, so let’s say a knowledge base, some sort of huge database that is well organised, so that when you receive an email next time, for example, well you don’t have to create the itinerary, the email reads how long people are intending to come, who they are, their background, does a LinkedIn search, bang, reads the past itinerary, prepares it, does it in a nice doc or word document, whatever way you do it, prepares the draft response to the itinerary in your inbox, and done. Before you even know there is a request in your inbox for such an itinerary. So that’s what I call being below the iceberg, an AI query, means AI can query all the great goodness that you’ve been doing as an organisation over the years and help you maximise the efficiency of it.

[11:35] So that’s a concept that is probably more technical and that we see staff and tourism organisation are starting to realise this could exist but they don’t know how to go about it. So our number one step normally is we tell them make sure everything that you do is in a digitised format.

[11:57] Liz Ward: And we’re pushing our clients with that, aren’t we? Like every meeting we have online we are finding a way to either at best transcribe that meeting is what we want. We can use that within a client’s wiki that we’ve created that is protected and private and gives us an amazing piece of context and detail information along with everything else about the business and their current AI and digital capabilities. That gives us an amazing client experience and that’s just one example of us making our content queryable and being able to do better work. And increasingly when I’m sitting coaching operators they are now starting to ask the question how do I use AI to improve the consumer customer process?

[12:46] How can I reduce the time and get better outcomes for the customer when somebody is emailing with questions or phoning and the staff are not really equipped to be able to optimise every part of that conversation and go back. So that’s the kind of work that we’re seeing starting to evolve and the adoption is starting to creep up which is fantastic because the awareness is there. Let’s move on. I think one of the most significant things in the travel purchase cycle in the last few months in terms of change that we’ve seen is this now it’s happening, we’re speaking about it for the last 18 months, it is an absolute collapse of the purchase cycle down into not only the search, that discoverability, but the booking, the actual transaction happening in one session. Now we’ve seen ChatGPT introduce online shopping basically.

[13:50] You can go in there and you see the products come up yet they backed away. They announced not long after they released the shopping that they weren’t going to go there for travel. They were going to leave that to the OTAs with the apps that they build for use within ChatGPT. Too complex, too many variables, yet not long after that Google’s jumped on it and said oh no we’re doing it, we’re going to have the purchase experience inside Google. Now this is absolutely massive in terms of changes to the travel purchase cycle and we’ll be responsive to so many players, not just DMOs.

[14:33] Fabienne Wintle: I think Google should have just changed their name. It would have made it easier for people to understand that this is the full 360 change. Like Google as we know it does not exist anymore. People, it’s a new thing, maybe call it Gemini. So what does that mean? I mean, it does make sense, doesn’t it, Liz? Like Google Business Profiles, they’ve had forever. ChatGPT…

[14:55] Liz Ward: It’s one of their greatest strengths is knowing everything about businesses and all those audits that they do that upset some of our clients so much because they got locked out of being able to respond to reviews or update their Google business listing because they had something amiss in there. So the naming, how they’d name their business compared to what their genuine business name is, that kind of thing or a mismatch on address. That’s Google’s way of making sure that they’ve got the most reliable accurate information about a business. They’re in a perfect position and they’ve got all the OTA relationships. They stitch those up as well.

[15:34] Fabienne Wintle: And they know us. They know…

[15:36] Liz Ward: Yeah. Yeah, and they know the consumer. So as a consumer, if I give Google access to be able to know more about me and authorise it to be able to do these purchases, I will be able to buy anything in a single session.

[15:55] Fabienne Wintle: Google created Gmail, right? Who does not have a Gmail? That’s how they know everything about you. They got you at the start. And I guess to really understand the strength of this, I’ll use ChatGPT as an example because I’m not super happy with the way Gemini works when I’m driving. But let me give you a story. I do a lot of driving. I like talking to AI when I’m driving because obviously I talk to it. I can’t type when I’m driving. So this gets you to start understanding how AI agents work, right? Let’s say when you’re talking to ChatGPT, it’s as if you basically outsource the work that you want to book a holiday to someone else. Like The Devil Wears Prada. Always imagine, you know, the assistant running behind you. So that’s your idea.

[16:46] Liz Ward: The reason this is important, that is because the purchase within Google and ChatGPT, it’s an agent doing the work. And that’s why we just keep going, explain the concept of agents.

[17:01] Fabienne Wintle: So let’s say, okay, imagine I outsource something to my assistant and she’ll do all the work for me. She knows me, just like Google knows me, right? She knows what I like. So I just say, okay, assistant, agent, go and book me a holiday. She comes back to me, she’s booked it. I’ve trusted her because she knows me. Now when I’m driving, currently I’m using ChatGPT because the voice of Gemini, I haven’t fine-tuned it yet enough. I found it a bit slow. But I’m talking to ChatGPT about where shall I go. It comes back to me. It tells me where I should go. Imagine now when the booking, the shopping cart has turned on Google, it would have locked it or maybe put it in my shopping bag and I would just have to verify and check out.

[17:43] I haven’t actually gone, read any information myself. So that’s what the experience of an agent, a travel agent, a PA is. So that’s what’s really important to understand for DMOs and how this has changed the Google we know forever. And Google, having the Google business profile means that, and it’s going to be much more reliable than any directory that a DMO might have on their side because it’s much richer. That’s changed the game for DMOs. So DMOs were the middleman before, but now not only AI has made it easier to access the information directly, but no one goes to the site. The agent provides the information, especially Google that knows everything about us and can close the deal. So DMOs have to think, the way that we were displaying content about the region, that is not required anymore because that’s been taken away from us.

[18:54] So what’s our role? How do we ensure that the perception that is given by our businesses about the destination on the LLMs is accurate?

[19:08] Liz Ward: And we’ve got some time, but not a lot of time. So we know that at the moment, the adoption of say fully autonomous agents booking for people, it’s very low at the moment. But if you consider that the investment that’s going on in AI technology and agentic technology, it is going to move very, very quickly. 80% of travel agencies, the OTAs are planning to deploy agentic booking at scale. So it really is urgent for the DMOs to understand what their roadmap is, to be able to be queryable, to be crawlable, and that carves out a refreshed value proposition for them. The Compass model that we have in the DMO AI playbook, it gives a very practical flow of what a DMO can do.

[20:06] And one of the most important pieces, I think even though it’s not technology based, the leadership that’s going to be required by any organisation in any industry is the most important part. So you don’t have to understand the technology, but you do need to engage the people who can make the decision and convincing the board is one, it’s one of the most important pieces in the Compass model. We have that as sort of the capital, the C, the first one, because that is what’s going to make the difference to get the decision makers to understand what’s happening.

[20:49] Fabienne Wintle: An easy way is give them a phone with ChatGPT and then they’ll quickly realise that making, I mean the content is really important still from a DMO, but the way it displays on the website is less. Making sure that that content is available to agents and think of talking to ChatGPT as ChatGPT being your agent is an easy way to visualise that. Making sure that the agent can access that content one way or another, is it going to retrieve it from your website or is it going to use an MCP which is basically an entry point into the knowledge base of the region to make that available to other agents.

[21:39] Liz Ward: Absolutely. We’ve outlined, there’s a couple of case studies between Toronto and Brand USA, just talking about what they’ve done to really lead on AI adoption and to get the right skills into their organisations and how to move forward. But I don’t think we need to think, oh we’ve got to be a great big Brand USA to do this. It’s any size DMO. If you value the local knowledge and the member relationships and all those, as you say, for decades of know-how about how to communicate the best things about your destination, then you’ve got the core asset that you need.

[22:24] Fabienne Wintle: And you digitise it.

[22:26] Liz Ward: Yeah, digitise it and do it, get the foundations down and then add the tools in incremental steps. Because once the digitised content is there, then they can really grow as an AI-enabled organisation. So I think it’s not doom and gloom. I think we can afford to be very optimistic because the value of this content asset, the information, the knowledge that sits in the DMOs is just an extraordinary asset. So I think we should be really encouraged by that. I encourage people, we’ll close it out here and I really encourage listeners to go and have a look at the AI playbook, which you can find on the tourismtribe.com website at compass-ai-playbook-dmo and that will be in the notes as well.

[23:24] And please make use of that, have a good discussion with your team and your board and we’re also happy at any time to do a quick debrief on the playbook with you and help you understand what your first step could be.

[23:42] Fabienne Wintle: Thanks, Liz, and see you next time. We’ll go into more details about the crawlable versus queryable.

[23:50] Liz Ward: Let’s do it. Okay, thanks for listening.

[23:54] Fabienne Wintle: Bye, everyone.

Free AI Readiness Score

How well does AI read your website?

Get your grade, a score out of 100 and your top fixes, by email. Free for tourism operators. No login and nothing to install.

Like what you hear?

Get access to all our training, coaching, and tools built for tourism operators who want to grow with confidence.

Explore membership