Speaking at WiT Singapore 2026, Etraveli Group CEO Mathias Hedlund said AI or large models have not yet solved the complete challenge of the flight back end, from procurement, pricing and fulfillment to service, and that this is precisely the business the company has bet on for twenty years.
Why recommended:The Etraveli CEO pulls the focus of competition from the AI front end back to the flight back end, and gives specific operating terms for the B2B platform's modularization and anti-fraud capabilities.
Meta launched its personal AI agent Muse in the United States on September 8, and Zuckerberg listed travel as one of its main use cases; within about two weeks, the market value of travel aggregation platforms evaporated by about $54 billion, and Booking and Expedia each fell more than 7% over two trading days.
Why recommended:The article sorts out the positioning choices and changes in bargaining power of OTAs, suppliers and distribution channels after AI agents become a new entry point for travel booking.
The article proposes that in the fourth-generation approval structure, the entry point becomes a natural conversational Agent entry, with expense control agents and TMC agents running in parallel, both invoked by an orchestration layer, and administrative control ownership is no longer a one-time selection. The author argues that expense control engines are designed for procurement across all categories and can only perform compliance validation, whereas TMCs hold real-time inventory, enterprise-exclusive closed-loop data, and offline fulfillment, enabling them to generate optimal solutions and make reasonableness judgments. The article also presents two yardsticks for the AI-ification of TMC services, namely input compression degree and decision convergence degree, and points out that TMCs need to open solution services to client enterprises' orchestration layers.
Why recommended:The article breaks down the division of labor between expense control and TMCs under the Agent orchestration layer into two layers: rule validation and solution generation, based on which readers can judge changes in the selection logic for travel procurement.