World-leading CX

Passengers judge support by how hard it is to get a request resolved. A simple flight change can mean failed website steps, phone menus, waiting and repeated explanations. Each step adds effort without completing the change. Even when the change is eventually made, the passenger remembers how hard it was.

Other priorities are put ahead of support. Investment goes into the flight, booking changes are restricted to avoid mistakes and spending on agents is limited. Each decision has a reason, but together they leave passengers with websites that do too little and long queues for the requests those websites cannot handle.

Customers want the right outcome without unnecessary waiting. Speed means reaching a resolution sooner. Quality means the request is resolved correctly and completely.

ChatGPT, Claude, Grok and Gemini have shown how capable conversational AI has become. A working self-service UI is far better than being forced to call, navigate phone menus and wait on hold. But learning each airline’s menus, completing forms and correcting validation errors puts unnecessary work on the traveler.

New technology is usually introduced one use case at a time: start small, test it, then repeat. But turning each use case into a separate workflow destroys much of the value of agentic AI. It constrains the model’s intelligence and makes customers fit the implementation, rather than resolving what they actually need.