The Shift Toward Agentic AI in Travel

Over the past year, the travel industry has begun to move toward agentic AI booking, an emerging model in which autonomous systems are being developed to not only understand traveler intent but also to search, compare options, and execute bookings on behalf of users.

Google is testing agentic booking capabilities that will allow users to book flights and hotels directly through AI Mode. OpenAI has partnered with Expedia and Booking.com to integrate their services directly into ChatGPT, turning the AI into a conversational travel planning tool.

Industry analysis increasingly highlights a shift from AI recommendations to AI executions. The competitive frontier is no longer just suggesting options—but completing the booking and managing the trip lifecycle.

The Bottleneck Core: From Intent to Reliable Bookings

AI systems are becoming highly capable of understanding what travelers want. The harder problem lies in turning that intent into bookings travelers can trust.

When travelers no longer open dozens of apps to compare prices, a critical question emerges: Can we trust AI agents to book on our behalf — in our best interest?

Not missing the best possible deals. No price jumps at checkout. No hidden rules behind cheap fares.

To meet these expectations, AI agents need access to high-quality, reliable, and structured data — the same level of information a professional travel agent would rely on.

Today’s reality is very different. Travel data remains deeply fragmented across suppliers, formats, and transaction flows. Airfare data is a clear example. It is highly fragmented across EDIFACT, NDC, and airline proprietary APIs — each with different integration schemas and transaction workflows. This complexity multiplies when content is distributed through millions of distributor APIs, where search speed, system stability, price accuracy, and content completeness directly determines whether a booking succeeds or fails.

Significant work is still required to standardize, normalize, and optimize this data before it is truly ready for an agentic booking era.

What Agentic AI Needs from Travel Data: Completeness, Accuracy, and Context

To enable agentic AI systems to book travel on our behalf, complete, accurate, and context-rich travel data must form the foundation. It requires comprehensive content to be aggregated, normalized, and continuously optimized — so AI can move beyond recommendation and execute bookings with confidence and consistency.

Content completeness is the first requirement. In an agentic booking environment, AI must evaluate far more than base fares. It needs access to multi-source content that includes branded fares, ancillaries, fare rules, baggage allowances, and itinerary details, all structured in a unified data format. Only with this level of completeness can AI agents serve different traveler profiles — whether price-sensitive customers or those seeking premium options — while reducing post-booking friction. When change and refund policies, restrictions, and conditions are embedded into the booking logic from the start, transparency improves, and traveler confidence increases.

Price accuracy is equally critical — and increasingly challenging. As agentic AI becomes reality, search volumes are expected to surge dramatically. At the same time, airlines continue to enforce look-to-book (L2B) limits to protect their infrastructure from excessive load. This creates a structural tension: more searches, but less tolerance for inefficiency. In this context, intermediaries must deploy smarter caching layers and predictive technologies to ensure prices remain fresh and bookable. Without this, agentic bookings risk last-minute price changes at checkout — eroding trust and undermining the very experience AI is meant to improve.

Beyond completeness and accuracy, agentic AI also requires contextual depth — data that explains how flight options perform in real-world travel scenarios. This goes beyond fare availability to include the practical impact of fare rules, and the identification of high-risk transfers, family seating constraints, visa-sensitive routings, overnight layovers, and regional operational norms. In human-led booking, this judgment comes from experience — knowing which options look attractive on paper but often lead to post-booking friction. For agentic AI to replicate this decision-making, data must embed context drawn from historical transactions and operational outcomes. Without it, AI may optimize price and availability yet still fail to deliver journeys that travelers trust and feel comfortable booking.

This is where B2B intermediaries become essential enablers — operating behind the scenes to modernize and stabilize the travel supply layer that powers AI execution. By curating, structuring, and continuously optimizing travel content, intermediaries help ensure that AI agents operate on data that is not only complete and accurate but also executable in real-world conditions.

In practice, this shift is beginning to influence how some intermediaries think about strengthening their data layers in preparation for an agentic booking future. From a data completeness perspective, this involves aggregating multi-source flight content beyond base fares. PKFARE, for example, curates content from over 600 airlines, including branded fares, ancillaries, structured fare rules, itinerary details, and baggage information — all normalized into unified formats that make future AI execution more feasible. From an accuracy standpoint, PKFARE maintains a 95%+ price accuracy rate through smart caching and machine learning — a prerequisite for agentic travel booking is to ensure fares remain accurate and current as search intensity increases. And from a contextual depth perspective, PKFARE’s more than 11 years of flight distribution experience — combined with historical data such as flight delay and cancellation rates, transfer feasibility, and change and refund policy— provides insight into real-world booking patterns. This accumulated context can inform more advanced AI decision logic as the industry moves toward agentic booking, helping translate traveler intent into journeys people feel comfortable flying with.

Agentic AI Travel Succeeds when Data is Ready

The agentic booking era is no longer theoretical. AI agents will increasingly shape how travelers search, decide, and transact. But the success of this shift will not be determined by who builds the most compelling interface or the smartest prompt.

It will be determined by who can consistently turn intent into execution.

In this new landscape, data quality is no longer a backend consideration. It becomes the infrastructure of trust — the difference between AI that can suggest and AI that can truly book, manage, and support a real-world journey at scale.

Agentic AI may be reshaping travel. But only those with execution-ready data will be ready to lead it.