AI can inspire travel, but it still can't book it. That's about to change

Artificial intelligence is rapidly redefining how travel is discovered, compared, and purchased. Over the past 18 months, generative AI platforms such as ChatGPT, Perplexity, and Google's AI-powered search have become embedded in the earliest stages of the travel buying journey.

The promise is compelling: a single conversational interface that can inspire, plan, compare, and ultimately book a trip - from flights and hotels to car hire and experiences.

But there is a fundamental question the industry must now confront:

Is AI actually fit for purpose when it comes to direct travel bookings - or are we still solving only half the problem?

AI is reshaping the top of the funnel

There is no doubt that AI is already influencing how travellers research and plan trips.

Industry estimates suggest that between 20% and 35% of travellers now use AI tools during their travel planning process, whether for inspiration, itinerary building, or destination discovery. Among younger travellers, particularly Gen Z, that number is already significantly higher.

Google has also highlighted a broader behavioural shift, with more than 40% of younger travellers beginning their journey outside traditional search, turning instead to AI-driven or conversational interfaces. This shift is profound.

AI is no longer just a search tool - it is becoming a decision-making layer, shaping preferences before a traveller even reaches a booking platform.

The booking gap: Where AI still falls short

Despite this momentum, AI still struggles with the most critical part of the journey: the transaction itself.

Today, most AI-powered travel experiences stop short of booking. The user is given recommendations - sometimes even fully structured itineraries - but is then redirected to external websites to complete the purchase. This creates friction, duplication, and often abandonment. The root cause is structural.

Travel booking depends on live, dynamic inventory - pricing, availability, fare rules, restrictions - distributed across fragmented systems, including GDSs, supplier APIs, aggregators, and direct connections. 

Most AI models are not designed to access or process this data in real time.

As a result, AI today excels at planning, but remains limited in execution.

Accuracy, trust, and the limits of static data

Accuracy is another critical issue. Recent high-profile examples have shown AI-generated itineraries directing travellers to closed attractions, incorrect locations, or even non-existent experiences.

In travel, this is not a minor flaw - it is a trust issue.

Unlike other industries, travel decisions involve:

• High transaction values
• Fixed dates and times
• Real-world consequences when information is wrong

Without access to verified, structured, and real-time data, AI risks undermining the very confidence it aims to build.

A new layer is emerging: AI-Native travel infrastructure

What we are now seeing is the emergence of a new category of technology designed specifically for this challenge.

Behind the scenes, travel infrastructure is being rebuilt to support machine-to-machine interaction between AI agents and live travel supply.

This includes new frameworks that allow developers to:

• Connect AI applications directly to real-time travel inventory
• Enable conversational systems to query availability and pricing dynamically
• Return structured, bookable results within AI interfaces
• Generate transaction-ready outputs, including pre-populated checkout paths

These systems are not traditional APIs built for websites or mobile apps. They are designed for a different environment - one where AI agents, not humans, are the primary interface querying and interpreting data.

For developers, this represents a significant shift.

Instead of building booking flows around user interfaces, they are now building data layers that allow AI systems to understand what inventory exists, how to access it, and how to convert intent into transaction.

In practical terms, this is what will enable AI to move beyond recommendation and into real booking capability.

Why travel is technically complex for AI

One of the reasons this transition is challenging is the inherent complexity of travel.

Travel is not a single product - it is a multi-layered transaction across multiple suppliers.

A single booking may involve:

• Flights with constantly changing fare classes
• Hotels with multiple room types and rate plans
• Car hire with location-specific inventory
• Activities with fixed availability and capacity

Each component introduces dependencies, constraints, and variability.

For AI to successfully handle this, it must operate on top of deep, structured, real-time data frameworks - not just generalised knowledge models.

This is why infrastructure, not just intelligence, is the key unlock.

From conversation to conversion

The next phase of AI in travel will be defined by its ability to close the gap between intent and transaction.

We are already seeing early signs of this shift.

Developers are beginning to build applications where:

• A user can ask for a hotel, flight, or package within a conversational interface
• The AI queries live inventory in real time
• Results are returned with accurate pricing and availability
• A booking can be completed without restarting the journey elsewhere

This transition is critical.

Without it, AI remains an inspiration tool. With it, AI becomes a distribution channel.

The role of developers in the AI booking ecosystem

One of the most important - and often overlooked - aspects of this transformation is the role of developers.

The next generation of travel booking experiences will not be built solely by traditional travel companies. They will also be created by:

• AI application developers
• Fintech platforms integrating travel into loyalty ecosystems
• Superapps and digital marketplaces
• Enterprise platforms embedding travel into broader user journeys

These developers need infrastructure that is:

• Easy to integrate
• Designed for AI interaction
• Capable of handling real-time, multi-vertical inventory
• Commercially viable, with built-in monetisation pathways

The platforms that provide this capability will effectively become the connective tissue between AI agents and the global travel supply chain.

Strategic implications for travel brands

For airlines, hotels, and intermediaries, this shift introduces a new reality.

AI is becoming a front-end layer for travel distribution.

Just as OTAs and metasearch platforms reshaped the industry over the past two decades, AI-driven interfaces will influence where and how bookings are made.

The key question is not whether AI will play a role - it already does.

The question is whether travel brands are prepared to:

• Make their inventory accessible to AI systems
• Ensure data is structured, accurate, and real-time
• Integrate into new distribution channels driven by AI

Those that do will benefit from increased visibility and conversion.

Those that do not risk being excluded from the decision-making process entirely.

So, is AI fit for purpose?

The answer is nuanced. AI is already highly effective at inspiration, discovery, and planning. It is reshaping the top of the funnel and influencing traveller behaviour in meaningful ways. But for direct bookings, it is still evolving. The missing piece has been infrastructure - not intelligence. That gap is now beginning to close.

A defining moment for travel distribution

What we are witnessing is the emergence of a new distribution paradigm.

AI will not replace existing booking channels overnight. But it will increasingly sit in front of them - filtering options, shaping intent, and directing transactions.

This is not incremental change. It is structural.

The travel industry has experienced similar inflection points before - the rise of online booking, the shift to mobile, the growth of metasearch. AI is the next one.

The difference this time is that the interface is no longer a website or an app. It is a conversation.

The companies that enable that conversation to turn into a transaction will define the next era of travel.