Destination Toronto & Visit Denver CEOs on the 5 Big AI Questions in Tourism

Welcome to the Destination AI newsletter from Matador Network, creator of the GuideGeek AI platform for DMO/CVBs.

Greg Oates, Director of AI Advocacy, Matador Network

A year from now, the travel and tourism sector might look back and thank Andrew Weir, president and CEO of Destination Toronto, for catalyzing how DMOs move forward with AI.

At PCMA Convening Leaders in January this year, I asked Weir what CEOs need to understand about AI at a high level. Numerous other CEOs have told me there’s a complete lack of AI education specifically designed for DMO chief executives.

His answer was first shared anonymously in our previous post, What Tourism CEOs Want to Know About AI.

“Tell me the five big questions about AI that I need to know, and what people are saying about them,” said Weir. “You don’t have to answer those questions but just tell me what I need to focus on. And tell me what people don’t want to talk about. What makes them squirm? Like, if AI is changing how people are searching for destinations, then how much should I stop paying Google?”

At the end of that post I suggested what those five questions might be based on all the CEO insights in there, and I’ve since been shopping them around to vet them.

A few weeks ago, for example, Richard Scharf, president and CEO at Visit Denver, suggested restructuring the questions to create a more logical hierarchy. Now, the first three deal with internal operations while the last two identify external, visitor-facing priorities.

Yesterday, I spoke again with Weir, as well as Paula Port, vice president of marketing at Destination Toronto, for their input on the five questions.

The goal here is to provide clear structure around AI platforms and processes for DMOs leading up to Matador Network’s collaboration with U.S. Travel at ESTO in Phoenix this fall. Together, we’ll be co-hosting an interactive, multi-station “AI Playground” educational activation during the three days of the conference.

Here are the five big AI questions. These will evolve but now we have a foundation to begin building consensus on how DMOs can capitalize on opportunities and navigate challenges with AI today.

1. How do we channel our staff’s enthusiasm for AI, who have wildly different perspectives and capabilities related to AI, to elevate our organization’s impact in our industry and community?
2. How do we create AI policy that serves and protects our organization but doesn’t suppress wonder and experimentation among staff?
3. How does AI benefit all our departments, versus just marketing, and how do we measure improvements in productivity, creativity and overall quality of work?
4. How do we redevelop our websites and content to rank higher in AI search?
5. How do we capitalize on conversational AI to influence travel purchase decisions and in-destination visitor behavior?

A growing number of DMO leaders say their teams embrace AI for the most part, and many are actively developing basic AI strategy. That’s a significant shift from a year ago. At the same time, there’s still varying degrees of cynicism and skepticism about AI in other organizations.

I asked Weir about his staff.

“I don’t feel like we’re dragging the team into this,” he said. “It’s really more a matter of how people are at such different places with AI. So our goal is figuring out how to channel all their knowledge and energy, and kind of align and unify that, without stifling anyone’s creativity.”

Scharf at Visit Denver emphasized that this question should be the first one, because successful AI integration begins with understanding your staff’s grasp of AI and their appetite for leaning into it.

“ChatGPT is growing like Zoom did a few years ago so we know we need to use it,” he said. “The first thing we did was survey staff to assess how they feel about AI, and how or if they’re using it. Then we started off with small bites like email. Now we think of AI as a toolkit, and we’re continuing to explore what tools are appropriate for different needs.”

Port added that AI integration across teams works best with both structured and self-directed learning.

“We started by making everything (ChatGPT Plus, Google Gemini) available to everyone, so they were given access to tools and resources, but a lot of this is really self-directed learning,” she said. “What we’re seeing now is people have taken advantage of that and moved forward with it. Some maybe a little bit further than others, so maybe we need to address that. I think we still need more use cases and more structured learning because we’re not all at the same level in terms of knowing and understanding how to put AI to use.” 

Destination Toronto was one of the first North American municipal DMOs to develop an AI policy. They began the process by pulling examples from larger government-operated tourism organizations including Destination Canada and Destination British Columbia.

“We set up that policy structure first in terms of what we wanted as an organization, and also being very clear that any use of AI was meant to drive business results,” said Weir. “I think we were able to get there pretty quickly because we didn’t create a series of regulations. We built a series of philosophies and principles to guide where we go. It was meant to put guardrails around protecting information, while at the same time, stimulate learning and sharing. Because we said to the team that you are free and encouraged to use AI, but there is an expectation that you will share what you’ve learned.”

Port explained further that, “We leveraged existing processes already in place from a content perspective in terms of bias, IT usage, etc. We have an assessment process for any sort of software application that we’re looking to use, so we did the same with AI.” 

(Note: The next GuideGeek Destination AI post will be with Kara Franker, president and CEO of Visit Florida Keys, who is also an attorney. We discussed AI policy at length.)

Scharf stated that integrating AI platforms and processes beyond marketing is “where all the work is happening now.” His organization is switching out regular stand up meetings with extended AI workshops and visiting speakers to learn more about AI tools for various use cases across departments. Leadership, he said, including himself, are present at these meetings so everyone is learning together.

Weir and Port both addressed this question, saying they need to first explore how Simpleview, Cvent and other tech partners are integrating AI functionality before looking at any new AI platforms. We discussed PCMA’s Project Spark, an AI-native event development tool that includes the new Destinaitor destination sourcing platform. DMO leadership and sales teams are exploring Spark to understand how clients are using new AI software to evaluate, compare and source destinations for business events.

Others, like Kara Franker mentioned above, are looking for new AI solutions to optimize their HR and accounting departments, but she said many of the available options presently are only enterprise grade.

Something DMOs should really look at, and I’m not sure how/if many already are, but there is proven value in creating custom GPTs for sales teams. GPTs are like mini ChatGPTs designed for specific purposes that anyone with a ChatGPT Plus account can create.

For example, a sales rep can build a knowledge base inside a GPT by uploading sales materials, examples of successful bids, product information, pricing, team members, case studies, testimonials and other data. The next step is to develop internal GPT instructions to direct what the GPT does and what it prioritizes. Following that, the rep can then chat with the GPT and ask it to create a customized bid for a potential client that pulls all relevant source materials from the knowledge base that best supports the pitch.

I’m imagining the ability to create custom GPTs is going to soon be a standard and required capability for all sales teams. One of the earliest enterprise examples of this was Salesforce’s Einstein GPT, which has since morphed into Agentforce. We use them at Matador because of improved efficiency and quality. GPTs also provide a new option for creative brainstorming, which some salespeople appreciate who don’t identify as being super creative.

Increases in sales production can be clearly identified. However, there’s increasing conversation about tracking success in other departments. When AI delivers clear improvements in efficiency and saves people significant time, how do you measure that? One concern is that people will then just work a six hour day versus eight hours. The common thinking is that organizations will need to adjust productivity benchmarks, but this is a messy conversation. I don’t know enough to comment on that yet, but I wanted to put a pin in it because it’s something we’ll all need to look at eventually as AI scales in the workforce.

I was addressing Visit Denver’s Board a while back and one member suggested that DMOs won’t need websites soon because they’ll be replaced by a ChatGPT-like window. I’m surprised by how much this comes up.

For one thing, DMOs will always need some kind of online knowledge base for AI search queries. If anything, that will require more content and more comprehensive product information because AI search thrives more than anything on data structure, clarity and specificity. That said, the web user interface will likely evolve considerably.

Expect to hear a lot more about how DMOs are developing data lakes/warehouses and knowledge graphs to optimize data structure for AI search. For further context, here’s a Perplexity Pro response to the question: “Are tourism boards developing data lakes and warehouses to optimize their sites for AI search?”

Weir suggested that this conversation about the future of websites is not unlike what we went through 15 years ago, when people were lamenting the death of DMO sites because social media was deemed by some as the new web. More important, he asserted, is seeing how AI search is evolving and being prepared to adapt strategy.

“I think there is a push-pull here between how we lead and where customers are at,” said Weir. “We want to deliver more information through our AI assistant and AI search in general, but customers aren’t really there yet. I don’t know what the percentage is, but the majority of people are still using Google. They’re still beginning an interaction on the web by entering search terms in Google the traditional way. At some point, that’s going to flip, right? It’s going to be a few percent more each month or quarter, whatever that is. But how far ahead of it do we get? And how do we calibrate? We can try to lead on this, but you know, we also have to stay in step with where our customers are too.” 

Chat platforms for DMOs like GuideGeek AI have dashboards that clearly identify what visitors are asking about in the destination. That helps inform content and product development strategy, and the data is valuable for community engagement. AI chat platforms also operate 24/7 in 50+ languages, which helps DMO influence in-destination travel and spending for a much wider audience than ever before.

How to best capitalize on conversational AI, and ultimately conversational commerce, is an evolving topic because we’re still in the early days where travelers are learning they can search with large language models and trust them. ChatGPT has 400 million weekly users but ChatGPT only added full functioning search capability for all users a few months ago.

Consumers are just beginning to understand that they can now “talk” to a city.

To help accelerate understanding and adoption of AI chat among visitors, Port asserted that the onus is on DMOs to promote their AI chat robustly like any other channel.

Weir did that recently by highlighting Destination Toronto’s GuideGeek-powered “6ix” chat platform on LinkedIn, asking, “Ever had a chat directly with a city?” Weir was specifically promoting the organization’s WhatsApp platform, where people can now have a highly personalized, iterative conversation directly with Destination Toronto via the GuideGeek AI integration. The same functionality exists with the other Meta platforms, Facebook Messenger and Instagram.

A few hours after Weir’s post I received a Slack message from our CTO Stefan Klopp wondering what was going on in Toronto. WhatsApp volume was spiking, he said, and he was asking if Toronto had sent out a press release promoting 6ix, or something. No, it was just a quick CEO social media post.

Also, Tourism New Zealand just launched a new campaign in collaboration with “A Minecraft Movie” that integrates GuideGeek-powered prompt questions to trigger further engagement.

So, those are just a couple examples of how DMO/NTOs can drive AI chat engagement.

The question for Weir and Port though is about the types of questions people are asking in AI chat platforms.

Presently, a lot of the chat conversations are more utilitarian, related to queries about hotels, restaurants, attractions, events, directions, etc. So as conversational AI adoption increases, will DMOs focus more on developing content that answers those basic types of questions, like what hotel, restaurant or attraction is highly rated, and how to get there. Or, should DMOs continue to focus mainly on developing what is generally perceived as more inspirational destination and experiential content?

“The question then becomes about responding to inquiries versus the inspirational side,” explained Port. “Like, all of these articles, all of this content that we serve up, that is not in response to a question per se. It’s more packaged to inspire people. I think a lot of that’s going to have to exist outside our environments, maybe, whether it’s influencers, it’s your earned media strategy, it’s content partnerships, things like that. That’s where I think the inspiration will exist. But we really have to move to being more of the informative one, on our website or whatever a website is. But it needs to be very specific in answering questions. We already see that on GuideGeek.”

Weir summed up, saying, “I think this ends up relating more to, again, that question of where the customers are. Are we putting ourselves in a position to answer their questions and deliver the information in the way they’re looking for it, using AI and the data lake approach? Or, are we curators like we’ve always been, inspiring visitors with local experiences, or is the customer now the curator by virtue of how they’re asking questions today?”

These five takeaways are aligned with the five big AI questions.

1. AI adoption begins with talking to DMO staff to understand what the organization is working with. Love the idea about marrying structured and self-directed learning, and being intentional about that with staff. At some point, for AI adoption to scale and continually evolve on an ongoing basis, there needs to be AI leaders in any organization who pick up the torch and inspire their colleagues to lean into AI. Otherwise, things tend to plateau with the same people using the same AI models.
2. AI policy is intended to encourage staff to innovate with AI as much as it’s meant to protect organizations.
3. AI is the death of “I don’t know,” so it can benefit all staff in all departments. That in itself is enough ROI to validate initial investments in AI. The tools can provide a foundation of knowledge and direction for problem solving and strategic planning exponentially better than traditional search.
4. Some organizations, like Visit Greece who we work with, have phenomenal website data structure ideally optimized for AI search. Meanwhile, one CEO told me “Our website is a mess with outdated content and poor data structure. We need to rebuild from scratch with AI search top of mind.”
5. Chat is the new marketplace, and AI is eating the entire funnel including in-destination. Moreover, AI chat data provides DMO leaders with insights into what visitors are asking about. Although, I think as chat adoption scales mainstream, the conversations will evolve beyond asking for a good hotel or restaurant.

For information about Matador Network’s industry leading GuideGeek AI chat platform, visit guidegeek.com/destinations.

For any DMO executives, schedule a call with me to discuss AI strategy at calendly.com/greg-matador.

For details about how GuideGeek AI was developed, check out this Crunchbase story written by Matador CTO Stefan Klopp: Reinforcement Learning From Human Feedback Took Travel AI Tool To Near-Perfect Accuracy.