How Generative AI is transforming the tourism industry

AI market size

What is the market size of AI and its impact on global markets? Various sources propose data on the market size of the AI industry. According to data from Statista, the worldwide market is forecast to exceed USD1.84 trillion (approximately JPY276 trillion)¹ by 2030, with a compound annual growth rate (CAGR) of 35.5% for the period from 2024 to 2030. Reports also indicate that applying generative AI across all industries could have an economic impact of USD 2.6 to 4.4 trillion², which represents an economic boost of 15 to 40%. It could potentially raise global GDP by 7% (approximately USD7 trillion) over the next decade and enhance productivity by 1.5%.³

The rise of generative AI has further expanded the range of AI applications and is expected to exert a significant influence on the global economy.

AI applications in the tourism industry 

At the recent WiT Singapore conference⁴ , attendees learned that about 60% of Asia-Pacific travelers are using AI tools to research and book their travel destinations. The primary objectives for these travelers are to reduce the time spent on booking, secure the best deals, access reliable and useful information, and overcome language barriers.

The results of a different survey targeting the hotel industry revealed that 63% of respondents are deploying AI to enhance revenue management, with common usage including data analysis, pricing decisions, and tracking market trends and the comparative performance with competitors.

As an example, Recruit (Jalan)⁵ and Kayak⁶ have introduced AI chat platforms for travel booking that offer personalized options for destinations and accommodations aligned with user preferences. United Airlines⁷ uses AI models to provide timely and accurate information about flight disruptions, including delays, to improve customer service. Concur Travel⁸ offers an automated tool that leverages generative AI to process business travel expenses efficiently and accurately. As a consequence, AI is playing a pivotal role in not only serving travelers with customized experiences but also in boosting business efficiency and personalized communication through AI-driven chat platforms. Priceline is also integrating OpenAI's Advanced Voice Mode technology into its AI chatbot. Voice-based interfaces are known to improve user experience (UX) by simplifying searches and travel booking⁹. 

Present approaches to employing AI can be broadly categorized into the following three areas:

(1) Personalization:

• Utilizing historical reservation data enhances personalized recommendations, resulting in greater satisfaction for travelers.

(2) Automation:

• Enhanced operational efficiency and productivity in the tourism industry
• Automated alignment of services and products with consumer needs, along with automated reservations

(3) Advances in communication:

• Improved UX with the introduction of chatbots, including automated translation, to replace lists with options.
• Evolving communication channels using digital avatars and other technologies, including extended reality (XR)

AI has its limitations

The advent of generative AI sees numerous companies aiming to deliver personalized information to travelers. However, without access to data, AI systems find it challenging to present viable solutions. In addition, unless a significant volume of individual preference data is available, the techniques for personalization are similar to those used previously, and typically involve providing information based on inferred preferences from clustering within large datasets.

The volume and quality (accuracy, recency) of data are crucial

A prevalent issue with generative AI is the reliability of its output. For instance, when trying to obtain travel information using the free version of the ChatGPT service, a traveler may occasionally encounter responses that are entirely irrelevant or illogical. This is due to the limitations of the learning data, which can lead to responses that lack the most up-to-date information or accuracy. As a consequence, numerous businesses are integrating generative AI with their proprietary datasets and leveraging the capabilities of generative AI interfaces to improve the accuracy of AI responses. In essence, while the volume of data is important, the reliability and recency of that data become critical. It is important to understand that merely addressing issues using generative AI does not guarantee an automatic resolution.¹⁰

Accumulating sufficient data from various domains is essential for developing effective AI-driven personalization

The benefits of "personalization" are highlighted as an advantage of employing AI. As mentioned, the key to personalization lies in possessing large amounts of data about individual preferences and behavioral histories. In other words, collecting personalized data related to reservations and activities during travel is challenging. This is especially true when travel is infrequent. In 2022, the average number of domestic overnight trips in Japan stood at 1.86 trips per Japanese resident, while the average number of domestic day trips was 1.48 trips per resident.¹¹ Consequently, personalizing services to match individual preferences and tastes can be challenging without maintaining historical data.^12,13   In addition, when using past data, we cannot exclude that a traveler's preferences may have changed over time.

To further personalize services, gaining deeper insights into traveler preferences is essential. However, a significant question remains: Can a service claim to be personalized if it relies solely on data from travel bookings and related information?

How to collect data

As we have observed, effective AI utilization requires sufficient data and collecting information on activities beyond travel can uncover traveler preferences at a given time. This leads to more personalized recommendations.

As an example, Google can extract from its search engine personal data that includes but is not limited to travel history: It is conceivable that they could access a vast amount of data about individual preferences and behaviors.

In recent times, our behavior has increasingly revolved around mobile devices. This suggests that successful integration of the information stored on smartphones may reveal the persona of each traveler, leading to enhanced personalized services.

In these cases, it would be difficult even for large platform developers to separate travel-related information from the remainder of the data that they also possess. As a consequence, companies are both incentivizing increased use of customer loyalty programs and are issuing credit cards to capture additional sources of consumer data.

However, acquiring more personalized data using the same strategies as major tech players is difficult for the tourism-related industry, which consists of many small and medium-sized enterprises. These businesses find themselves at a disadvantage to major operators intent on delivering personalized services with the data that they possess. As a result, efforts to gather data on traveler behavior in the region have focused on linking and integrating IDs.

Vendor relationship management (VRM) approaches are the complete opposite of traditional methods

In terms of data utilization, we might typically envision a model where businesses collect data from individuals and then provide information tailored to those preferences. This is commonly known as customer relationship management (CRM). Until now, cookies have been pivotal in tracking user behavior to collect information. However, with the movement to phase out third-party cookies, which enable tracking across multiple sites, there is a growing shift toward direct collection of first-party data.¹⁴ Additionally, there is increasing collaboration and sharing of data between businesses, facilitating the use of first-party data collected by other companies.

How does tourism marketing influence perceptions of a destination? Kinosaki Onsen, a hot spring resort in Hyogo Prefecture, is an example of how information on about accommodation is aggregated by region and used by destination marketing organizations (DMOs).¹⁵ Kesennuma City in Miyagi Prefecture also serves as an example, having introduced a point card system with integrated IDs to centrally manage tourist data collected across the region.¹⁶

Efforts to aggregate data through regional collaboration have begun in various parts of Japan. However, progress is often slow due to resistance from local companies that are concerned about sharing their proprietary data with other businesses.¹⁷ As just illustrated, the use of AI relies heavily on data as a critical element. If it is difficult for businesses to collaborate in collecting data, could they instead have customers directly provide the data? This concept, known as VRM, can be seen as the reciprocal of CRM.¹⁸

There is often debate about "Who owns the data?" In the past, the standard approach for businesses was to gather and make use of customer data. In parallel, customer transaction information is retained as part of the customer's data record. As an example, when shopping at a store, information about purchases is accumulated by the store, while the data is also provided to the customer in the form of a receipt. There are services that manage the customer’s data in the form of a digital wallet, consisting of daily information that customers manage themselves. Credit card statements, travel itineraries, and detailed information about an individual's health, medical, and nursing care recorded at multiple medical institutions are also managed at the individual level with tools like personal health records (PHRs). While there are multiple ways to store and handle this data, it is all gathered and maintained as personal information.

The core concept of VRM says that individuals can receive optimal service by sharing self-collected data with businesses that provide the services that they want.¹⁹

Even when a single business does collect data, it can only know the part of the customer which interacts with the services it provides. VRM involves receiving data directly from customers in an approach which fosters a deeper and more holistic insight into customer needs and preferences, as opposed to forming an incomplete view based on piecemeal information. Currently, a variety of transactions are being conducted primarily through smartphones. In various discussions surrounding AI, there is a notion that, in the future, each individual may have their own AI agent, which would interact with other AI agents from different businesses to facilitate transactions. There may be a time when AI agents provide service providers with vast amounts of personal data accumulated on an individual’s smartphone. This would allow customers to receive services more suited to their preferences and personalized information.

Generative AI needs people who possess the skills to interpret data and derive meaningful insights from it

In discussions on Digital Transformation (DX) initiatives, there is a growing emphasis on the need for talent such as data scientists to serve as DX professionals. Employing individuals with expertise in sophisticated statistical analysis would enable businesses to derive a broad spectrum of meaningful insights. Nevertheless, without complete datasets available, opportunities to utilize advanced data analysis will be limited.

With the advent of generative AI, the emphasis often lies on offering personalized customer experiences. However, it's equally crucial to leverage this technology to increase the productivity and enhance the management of tourism-related businesses and destinations.

Previously, data analysis posed a challenge for people lacking knowledge of Excel basics, advanced applications and statistical analysis. The emergence of generative AI has revolutionized data analysis. By clearly specifying the analyses that you need and supplying the relevant data, generative AI can execute statistical evaluations to deliver the result. At the forefront of this revolution is Generative AI’s transformative technology that can enhance data visualization through the use of graphs and other visual representations.

These developments suggest that it is now possible to boost productivity and improve management without necessarily possessing knowledge of the analytical methods. Of course, learning about analytical methods is important; however, perhaps even more crucial is the ability to interpret the results of those analyses, to understand what the numbers signify, and to make sense of the data. This skillset may well define the profile of sought after talent in the generative AI era.

Generative AI has the potential to be a transformative force for tourism business models

As the discussion on generative AI progresses, many instances underscore its potential for personalization, streamlining processes, and enhancing interactions. However, does this encompass the complete range of AI's abilities? Past data can be a valuable source for tailoring recommendations because it provides insights into a traveler's preferences and behaviors. However, it does not guarantee a traveler's satisfaction.

Travelers do not always choose their journeys based on past travels. Changes in external factors and shifts in personal circumstances and preferences over time should also be considered. Travelers best understand their own preferences and needs.

If we compare the business models of the tourism industry from the offline era to those of the online era, we can observe several distinct features.

In the offline era, travel agencies specialized in understanding the preferences and requirements of travelers and, in response, curated a selection of accommodations and experiences tailored to meet those specific preferences. Since the dawn of the online era, people have started to make travel choices through online travel agents (OTAs). In doing so, travelers have moved away from expressing their preferences as they did in the past: they now select accommodation and experiences from OTA listings to make their travel decisions. In this regard, the focus is on suppliers, including accommodation and experience providers. In the past, whether operating offline or online, suppliers typically provided inventory to travel agencies, agents and OTAs, and then relied on them to refer customers. There is a widespread belief that suppliers often struggle to attract customers directly, primarily because of resource constraints.

Travelers now select accommodation and experiences that match their preferences, including price, from OTA listings. But can they truly find what they desire in these listings? Travelers may frequently experience uncertainty and find it difficult to articulate their choice of destination. For example, instead of choosing a specific destination, a traveler might list desired activities, such as visiting a location to enjoy forest bathing about an hour and a half from the city center, having an evening BBQ with friends, or exploring water activities.

If we could visualize the personal tastes of this traveler, suppliers would have the opportunity to see and meet those preferences, offering a novel and unparalleled value in the travel experience.

This would lead to the empowerment of suppliers. Lodging and experience providers are facing challenges in conveying their appeal to tourists via OTAs and similar platforms. If the preferences of travelers are clear, suppliers have the opportunity to engage with travelers directly for the first time by providing services that cater to those preferences, with the exception of direct bookings with travel agencies.

Generative AI has significantly improved user experience (UX), potentially leading to an evolution in the role of OTAs as these technologies advance.²⁰

Leveraging generative AI capabilities

Generative AI offers many opportunities but also requires a significant amount of high-quality data. The tourism sector, predominantly made up of small and medium-sized businesses, faces challenges in delivering personalized services to tourists by relying solely on in-house data. In the future, enhancing UX through collaboration between local businesses and services outside their own core offerings will gain increasing importance.

In addition, by leveraging generative AI in-house, employees can develop the ability to analyze data without having to master advanced statistical analysis methods, which leads to improved productivity and more sophisticated management.

The development of generative AI has the potential to significantly enhance the value of travel experiences through major improvements in UX. It can also empower suppliers, including accommodation and experience operators. As the technology progresses, there is a growing need to adopt and integrate generative AI into business practices while continuing to conduct research.