6 minutes read
What this article talks about
- Why AI struggles in high‑stakes travel scenarios where disruption, emotion, compliance and incomplete data require human judgement.
- Why data quality and infrastructure matter for accurate decisions, regulatory trust and reliable automation at scale.
- How AI and humans work together in travel with AI handling routine tasks while people guide complex, subjective or high‑risk situations.
Travelport’s Senior Director of Customer Success & Training Delivery Claire Osborne says trust in AI is tested when the risks increase
AI is great for destination discovery and travel planning. It excels in low-risk, high-volume operational tasks. But it doesn’t cope so well in high-stakes scenarios such as delays, disruption and rebooking—especially when complex security and compliance requirements apply.
You don’t need to read Harvard Business Review to know that competitive advantage increasingly comes down to how effectively a business can solve complex problems—and how quickly it can adapt when problems change, or new ones emerge. I see that every day. My job is to help Travelport customers succeed by addressing complex business needs and adapting to new challenges and opportunities. AI will play a major role in that.
Under pressure
AI performs well as a decision-support tool: when it has access to structured data, clear inputs, and repeatable outcomes. It works best when there are defined rules around the required actions; for example, airline filing data and the parameters governing its use. But human oversight and judgement become critical when emotional responses come into play, or when the available data is insufficient to drive a fully automated decision. Different travelers want different things, and today AI cannot fully infer those preferences without far deeper personalization.
Take schedule changes. It may be logical to rebook a traveler on the next available connecting flight. Yet that traveler may prefer a different route or to travel the following day, based on factors the data cannot capture (“I can’t bear connecting through JFK”). In many situations—especially during disruption—human empathy is essential to reassure and guide. AI can do the heavy lifting in presenting options. But when the answer to “What’s best?” is subjective; human experience and judgement remain vital. Over time, AI will improve, with better customer profiling and historical data enabling more nuanced decision support.
And what about regulatory frameworks and compliance constraints? Can AI be trusted in such a complex, high-risk environment, with legal, financial, and reputational implications? Perhaps in time—but only with the right foundations in place.
By that, I mean a robust infrastructure built on accurate, validated data: a platform that is not only AI-ready but also designed to address the complex and evolving compliance needs of travel. A key aspect of Travelport’s platform is meeting regulatory needs, whether that’s financial standards like PCI, data protection such as GDPR, or frameworks set by IATA. Our fares and tools also include guarantee policies, so when customers use Automated Exchanges, they can trust that transactions are protected.
Data quality is key
Effective AI depends on high-quality, standardized, and structured data. Poor or inconsistent data limits automation and increases the need for human intervention. Our platform already applies machine learning across data elements such as bookability, helping ensure customers receive the most relevant results. In many cases where customers can choose from more than one GDS and have multiple booking options, we see increased selection of the Travelport platform. There’s no stronger validation than an uplift in bookings.
A changing landscape
Operational frameworks are evolving as AI becomes more capable. There’s huge adoption of a ‘chat first’ approach, and using AI to automate processes where there is sufficient precedent and reliable data.
Well-designed chatbots provide fast, accurate responses, resolve many routine queries, and escalate seamlessly to human agents when needed. This is about giving travelers greater control—automating straightforward tasks such as changing a seat without multiple steps or quickly checking the cost of altering a flight. AI improves speed and efficiency. In a world where time is increasingly valuable, that translates directly into better service. It also allows organizations to deploy human expertise where it matters most—applying emotional intelligence in more complex scenarios. However, poor AI implementation, such as basic FAQ-driven chatbots, can quickly lead to frustration, repetitive loops, and negative brand perception.
THE CUSTOMER VIEW: AMERICAN AIRLINES
AI has the potential to create immense value, but don’t become a hammer that is looking for a nail - rather build the infrastructure to solve the lowest-hanging-fruit first
Marcial Lapp, VP of Revenue Engineering at American Airlines
It’s the next step in a technology progression for airlines—from web to mobile into AI—which means incremental adoption across customer and operational processes.” The most significant near-term impact is in efficiency and scalability: “AI delivers the most value today by automating high-volume, contextualized workflows such as customer service claims handling, improving speed, consistency and cost efficiency while allowing human personnel to focus on more complex tasks.” Tomorrow, we expect AI tools to press into the more complex scenarios that combine a plethora of travel options (itineraries, products, ancillaries, competitors) along-side with customer-preferences to augment trip-planning and execution.
Marcial also highlights current limitations around complexity, data readiness, and risk sensitivity: “Data is the key to making your AI processes and agents successful.” As a service-provider, control of said data is key, and American Airlines is investing in modernizing platforms and processes to more seamlessly allow AI to function. This investment starts with data systems, and your API-first-now-turned-MCP strategy to support task execution.
AI scales decisions but people still define them
AI is already helping our products evolve, enabling more self-service options and improving the speed and relevance of content delivery. But it remains essential that users can step in and make decisions where AI cannot, and do so seamlessly. As AI proves itself in more routine areas, business and user confidence will grow. We’ll see AI increasingly connecting the dots—for example, recognizing that a flight change may require updates to hotel bookings, car hire, transfers, and even restaurant reservations. Delivering this requires the right data, integration, and connectivity across systems. Humans need to define the parameters and—as my colleague Andrew Jordan wrote in the previous blog in this series—travel sellers really need to focus on adapting their culture, systems and processes
AI will enable deeper segmentation and more tailored offers across both leisure and corporate travel. Systems will increasingly surface and prioritize the most relevant options, with recommendations based on real traveler behavior.
In short, we at Travelport are moving towards greater personalization, precision, and automation at scale, with AI supporting more of the end-to-end journey. We’re building the infrastructure and products to make this possible. Ultimately, Customer Success at Travelport is about partnering with our customers to help them get maximum value from our platform, in which AI will play an increasingly important role, from supporting MyTravelport interactions to providing travel consultants with consistently relevant, bookable results at the right price point.
This is the second in a blog series exploring how AI is transforming travel—and how the businesses that succeed will be those with the infrastructure, platforms and expertise to make AI work at scale.
The next blog describes how underlying infrastructure, not AI models, will determine who wins.