Travelport, Anthropic and Cognizant partnership

Building the infrastructure for AI-era travel

Today we are announcing a strategic partnership with Anthropic and Cognizant. We are putting advanced AI directly into Travelport’s platform - and into how we build, test and run the software behind it. Three companies, one shared goal: make AI work for real travel transactions, not just trip inspiration.

I want to take a few minutes to explain why we are doing this, why we are doing it with these two partners specifically, and what it actually changes - for our customers, and for the wider industry.

Why now

Travel is in the middle of one of those rare shifts where the underlying behaviour is changing faster than the underlying technology. Travellers are increasingly searching with natural language. They are talking to AI assistants. They are skipping traditional booking flows entirely and expecting a system that understands intent, context and preference - not just keywords and date filters.

That is not a future trend. It is happening now, and the pace is accelerating.

The catch is that almost none of the systems that actually power travel bookings were built for this. Most distribution infrastructure assumes the traveller knows exactly what they want, types it in cleanly, and clicks through a deterministic flow. Drop conversational AI on top of that, and you can do a beautiful job of inspiring a traveller - right up until the moment they want to actually book, hold, change or service the trip. Then everything falls apart.

That gap is the problem we are solving.

What we are building

Travelport’s TripServices platform is the cloud-native API layer we have been building for the last several years. It already provides unified access to airline, hotel and other content through modern APIs designed for orchestration and AI-ready workflows. It was built deliberately to handle both ends of the search spectrum - the traveller who knows exactly which flight they want, and the traveller who is still figuring it out.

This partnership puts Claude - Anthropic’s model family - directly inside that platform, and directly inside our engineering organisation. It is paired with Cognizant’s enterprise AI delivery capability so we can move at real pace rather than treating this as a multi-year proof of concept.

The combination matters. A great model without solid infrastructure is just a demo. Solid infrastructure without a great model is a missed opportunity. Pair them properly, and you have something the industry has been waiting for: AI-driven discovery directly connected to live, bookable, serviceable travel content.

Why Anthropic, and why Cognizant

Choosing partners is not a beauty parade. It is about picking the right people for the specific problem in front of you.

Anthropic brings two things we needed. The first is Claude’s large-context reasoning - the ability to handle the scale and complexity of a platform like ours without losing the thread. The second is the rigour of how Anthropic approaches enterprise AI: safety, reliability and predictability sit at the centre of the model, not bolted on afterwards. Travel is one of the most fragmented and high-stakes transaction environments in the world. We need AI that performs in that environment, not one that performs in a sandbox.

Cognizant brings delivery. A lot of AI initiatives stall not because the technology is wrong, but because the implementation runs out of road. Cognizant has the enterprise experience and operating muscle to take what we design and to help us embed Anthropic’s frontier models into everything from code design and delivery to maximising the value of AI in our products. That is the difference between announcing an initiative and delivering one.

Travelport brings the rest: the platform, the content, the distribution reach and the customer base. Without that, AI in travel is just a clever assistant with nothing real to do.

Where this shows up first

This work lands in two places and both matter.

The first is inside our own engineering organisation. Claude will change how we design, build, test, document and maintain software across the company. That means meaningfully faster delivery, better code quality, sharper testing - and, importantly, engineers spending more of their time on the interesting problems and less on the repetitive ones. The early signal here is strong. We are not theorising about productivity gains; we are seeing them.

The second is inside the Travelport platform itself, where AI capabilities become a native part of how airlines, agencies and OTAs retail and service travel. That is the part our customers will feel directly. They do not need to build their own AI stack, they do not need to integrate a dozen point tools, and they do not need to wait for the rest of the industry to catch up. The intelligence sits inside the platform they already use.

Both tracks reinforce each other. The faster we modernise how we build, the faster we can put new capabilities into the platform. The more capable the platform becomes, the more our customers can do with it.

What this means for the industry

The travel industry does not get many genuine inflection points. Online booking was one. NDC, over the last decade, has been another. AI is the next, and it is moving faster than either of the previous two.

Airlines need smarter retailing that responds to traveller intent, not just transactions. Agencies and OTAs need to reach travellers who are searching through AI tools rather than traditional booking engines. Developers and technology partners need APIs that speak the language of AI workflows rather than legacy formats.

None of that gets built by accident. It gets built deliberately, by people who understand both the AI side and the travel side and are willing to commit to the work. That is the role Travelport is choosing to play.

I have said before that Travelport is becoming the infrastructure layer for AI-driven travel commerce. Today’s announcement is a meaningful step in actually delivering on that - not a statement of intent.

What is next

We have a clear roadmap, and our team is energised. Customers will start seeing the impact in product capabilities quickly. Internally, the engineering uplift is already under way.

This is one of those rare moments where strategy, technology, partners and timing line up: we intend to make the most of it.