DALLAS — For the past decade, airports have tried out various 'smart' tools like dashboards, predictions, and alerts. Now, a new phase is emerging: agentic AI.
In an interview with International Airport Review, SITA Labs’ Jordi Valls explains that this shift means moving from assistive or predictive AI to systems that can weigh goals and limits, then suggest or carry out coordinated actions across teams and tools. Instead of just saying 'a delay is likely,' these systems can outline the trade-offs and recommend the best steps to recover.
Airports: Faster Disruption Recovery, Not full Autonomy
The article says that aiming for 'total airport autonomy' is not the right goal. Instead, partial autonomy in specific areas like handling disruptions and reallocating resources is more realistic, with humans still in charge. (International Airport Review)
The quickest returns mentioned are expected in tough, time-sensitive parts of airport operations, such as disruption recovery, turnaround coordination, and baggage issues. These are areas where there is lots of data, but decisions often fall apart under pressure.
A unified operational data system is essential. Having a shared and consistent source of truth for flight status, resources, passenger flows, and constraints. Without this, the warning is clear: agents will make decisions based on conflicting information, leading to unreliable automation and loss of trust from operators. This is especially risky in safety-critical settings.
Expedia: Agentic AI Is a Direct Competitive Threat
According to Skift, Expedia’s latest annual report clearly states that generative and agentic AI will make competition tougher. There is a risk that new companies could use AI to offer better or faster travel search, planning, and booking.
Skift also notes that Expedia is working to stay visible within agentic experiences, but is concerned about data scraping, liability, and the fact that its own agentic features for customers are still in the early stages.
Our Take
We believe that both airports and travel sellers face the same challenge: agentic AI benefits those with clean, connected operational and inventory data. Airports that do not unify their data will end up automating chaos.
Travel platforms that do not make their inventory easy for agents to use risk becoming just a hidden backend while the agent controls the customer relationship. The winners will not be those with the fanciest chat interface, but those who can turn scattered systems into reliable, trackable actions at scale.
What we'll see customer-facing platforms catering content to humans and AI agents alike. There's a proposed, AI-optimized standard for a text file placed at a website's root (llms.txt) that provides a structured, Markdown-based map of content for Large Language Models (LLMs).
This file serves as a curated roadmap, helping AI agents like ChatGPT, Claude, and Gemini quickly find, read, and interpret the site’s most important information. In practice, it points toward a future where you can share an itinerary with your device and an AI agent can complete the entire booking flow on your behalf.
Whether the agent uses Expedia is another question.


.webp)
.webp)
.webp)
.webp)
.webp)




.webp)