Scrolling through the 2025 travel and tech
trends content that dominate this time of year, very few seem to understand
that generative AI (genAI) is much bigger than a trend. GenAI is already having a fundamental
impact on the travel marketing, merchandising and selling value chain for
suppliers. It is taking over from legacy search, with conversational and
agentic AI driving this disruption.
As a result, Google's absolute stranglehold
is clearly under threat, probably more so than people realize.
Google’s AI will not be as dominant as its
search engine because we’re already seeing loads of large language models
(LLMs) in play in travel and hospitality, creating even more fragmentation than
we’re used to.
As genAI continues to grow, there will be
many more places to seek advice on where to travel. There will be agents everywhere.
This is where the opportunities for
innovators and investors lie. Not in finding ways to monetize a new world of
search no longer dominated by the need to optimize for Google; it’s about
orchestrating the data, helping companies to address tech debt, developing the
API services that connect different systems and stacks so that data and content
can be ingested by the LLMs and surfaced to the end user via a conversational
or agentic channel.
Push and pull factors driving the nails into the coffin of legacy search
Conversational and agentic generative AI, in
particular, is changing how people search. Queries fed to AI tend to be more
long-tail and granular than legacy searches, and this change is evidenced in
corporate and leisure travel. Thayer investment Groupize, a meetings and events tech specialist,
bought
a meetings space start-up, The Vendry, this June with the clear intent of
embedding The Vendry’s artificial intelligence-powered search capabilities into
the Groupize platform, giving planners, planning advisory services and travel management companies deep information about more
than 250,000 potential venues.
At the same time, a
generational shift is pulling users away from legacy search — more 18 to 24-year-olds are using
Instagram and TikTok than Google — but the changes to how users engage are ubiquitous and beyond
generation. So, there isn’t a ten-year window of opportunity for suppliers to
get ready. High–spending boomers are using genAI today in their travel
planning, and Phocuswright's recent Travelers and Tech data shows that nearly one in five (18%) travelers in the United States have already used genAI.
Ideas about where to go and what to do top
the use cases, confirming discovery and inspiration as the main drivers of
engagement with genAI.
Relevant here is that genAI sources and
surfaces its responses differently from legacy search, even if the queries are
in the same ballpark. Content has traditionally been indexed for Google, and
sellers and suppliers prioritized optimizing their content and data around
keywords, backlinks, UX, site speed, etc., so that they could appear on the
first page of Google and be visible and discoverable at the top of the funnel.
With genAI, conversational search is now
the front end, and there is no guarantee that a hotel, attraction, or online travel agency that
ranks highly on Google legacy search will be recommended by the conversational
AI or surfaced by the agents.
This fundamental change needs to be
addressed because we’re talking about an existential threat to the thing that
matters most — revenue. If people using genAI aren’t discovering you (because
you’re not discoverable by LLMs), they are not booking you.
And this threat is driving a positive
change in corporate culture at the C-suite level. The motivation for leadership
to get serious is different because businesses must pay attention to something
that could threaten a significant percentage of demand. CEOs and CFOs see revenues
dropping, so they ask why. Engineers say, ‘It’s because you've never given me
the budget to fix our tech stack,’ so
the CEOs are saying, ‘ok, let’s get our tech stack in order, let’s work on data
orchestration, let’s be more proactive….’
This change of tone syncs with the boost tech
enjoyed coming out of the pandemic when leadership was caught flatfooted and
lacking agility. The perception of tech
shifted from something in the wings to center stage, with many execs actively
seeking to engage directly with startups and entrepreneurs rather than ignore
their calls. Thayer’s hotel tech investments, such as Mews and Canary, benefit
from this C-suite mindset shift, winning new customers and raising more funds
to invest in a compounding cycle of tech innovation that leads to more customer
adoption. It’s the essential flywheel of
adoption and innovation our industry has lacked for generations.
Another positive indicator is that today’s
entrepreneurs have access to tools — low-cost computing power, GenAI coding, content
and image creation, open systems, APIs, and support networks — that enable
rapid experimentation and pour fuel onto the fire of their creativity.
These factors are responsible for unleashing
a whole new wave of spectacular opportunities for innovators and investors.
The genAI industry, travel and hallucinations
Travel is at the forefront of generative AI
because it's a complicated challenge to solve.
Almost every demo the high-profile model builders release includes
travel planning as a use case. But the
difficulty is that they too are having to work with the industry’s tech debt
when it comes to search and discovery and are bringing up a high number of
wrong responses and full-blown hallucinations. One report found that 90% of itineraries
generated by ChatGPT in response to a query contained at least one mistake.
Agentic and conversational genAI bots are
mining, organizing and summarizing data from sources that are not optimal for
AI, from sources designed for human interaction. The parameters of the LLM
response are dictated by the rules it has been programmed with, relating to the
data set it has been trained on. As things stand, genAI’s killer application
for now is information retrieval and summary, but in the future, it will be the
enablement of commerce in itself. Most are not trained to validate the information,
and as genAI’s influence progresses down the funnel toward booking, the
downside of these hallucinations gets more intense.
Many of these errors come from incomplete
data, not just bad data. Travel
suppliers, in particular, need new data management and orchestration playbooks,
having historically focused on static search strategies. The emerging world of
genAI search demands a new framework for data orchestration in terms of
granularity and depth. Information is more dynamic; new information is
generated, and existing data is updated. Entrepreneurs see a path to leverage
this demand for tech that cleanses data, keeps it fresh, ensures its accuracy,
and then feeds it to the LLMs correctly so that it can be surfaced to the end
user. Today, Thayer is collaborating with a stealthy start-up developing
technology for brands to take advantage of this emerging genAI “superpower.”
Case (study) in point
Thayer investment NLX is a conversational
AI specialist deeply invested in many aspects of the transition away from
legacy search. Its patent-pending proprietary technology humanizes the dialogue
flows so that the conversation between the AI and the end user comes across as
unscripted and responsive (with the tech allowing the seller to apply their own
business logic).
NLX has also addressed the fragmentation
issue by ensuring that its platform can work with whichever LLM (and NLP) the
seller or supplier is using.
And finally, it is grounded
with Google, meaning that any response it generates is cross-referenced and
fact-checked with Google search and AI, reducing the number of errors that we
highlighted earlier on.
Some supplier clients are using early iterations
of NLX Journey. The technology is built to be scalable, so when the travel
industry addresses its tech debt and legacy search architecture, travelers might
book their trips within a single conversation.
Conclusion
It’s evident that genAI is shaking
everything up, but my enthusiasm is not limited to genAI itself. With the
C-suite saying we need a modern tech stack so we are visible to the LLMs, other
innovation opportunities come to light. We can rethink loyalty schemes, get
serious about personalization, overhaul food and beverage tech and more.
This wave of genAI acceptance is so
disruptive and absolute that it will unlock the cleansing of the tech stack,
lead to the retirement of technical debt and give innovators the opportunity to
build great businesses. It will be possible to do things you couldn’t do in the
past, provided, of course, that the data foundations are in place.