The airline distribution industry has a habit of fighting the last war. For years, the central conflict was between airlines and GDSs: who controls the offer, who captures the margin, who owns the relationship with the traveler. NDC was the weapon airlines built to win that fight. Now, before that fight has even been resolved, the battlefield is shifting again.
Model Context Protocol is not, strictly speaking, a distribution technology. It is a protocol developed by Anthropic that allows AI agents to interface with external systems, APIs, and data sources without requiring custom integrations for each one. In travel, the implications are significant. An AI agent with MCP access does not follow a booking flow or depend on normalized interfaces. It reads, compares, and acts directly on airline content, in real time, in whatever format the airline provides. The booking process, as it has existed for decades, starts to dissolve.
Serko’s Johnny Thorsen put it bluntly at the UATP Airline Distribution 2026 conference: “MCP will eliminate NDC.” Sabre’s Gonzalo Jorge pushed back, arguing that NDC retains value in content normalization and that the two can coexist. Both positions are defensible. But the more consequential question is not whether NDC survives technically. It is whether NDC remains strategically relevant when the decision layer sits above it.
That is the shift that matters. NDC was designed to give airlines control over how their content is presented and priced across distribution channels. It relocated the intelligence of the offer from the GDS to the airline. MCP relocates it again, this time to the AI agent. If the agent decides which content to surface, how to compare it, and when to book, then the airline’s ability to shape that decision through NDC schema and offer management becomes secondary. What matters is whether the agent includes you, trusts your content, and ranks your offer favorably. That is a problem of AI model behavior and orchestration logic, not XML compliance.
The aggregator that is most legible to AI agents becomes the distribution infrastructure of the next era. NDC remains in the stack, but it is no longer the surface that matters. TPConnects made this structural bet explicit on April 7, launching MCP integration within its Iris platform and giving travel sellers a machine-readable interface to 60-plus airlines across NDC, GDS, and low-cost carrier channels through a single layer. The message was unambiguous: the position worth holding is not content access, it is AI-agent compatibility.
This is where the value chain analysis gets uncomfortable for any airline that invested heavily in NDC as a strategic differentiator. The premium that NDC was supposed to deliver, richer content, dynamic pricing, personalized offers, only materializes if those offers reach decision-makers in a form they can evaluate and act on. In a world where the decision-maker is increasingly an AI agent running on MCP, the airline needs to be legible to that agent. Airlines and aggregators that build MCP-compatible infrastructure early will not just have a technical advantage. They will have a positioning advantage at the exact moment the industry’s offer architecture is being rebuilt.
For global carriers with large technology budgets and mature NDC implementations, this is a manageable transition. American Airlines, Lufthansa, and their peers have the resources to layer MCP capability onto existing infrastructure and experiment with agent-facing distribution. Sabre launched its own MCP server in September 2025 to connect its APIs to AI agents. These players will adapt.
The situation is materially different across Africa. Africa risks completing the transition to NDC just as the industry begins to move beyond it.
Kenya Airways, Ethiopian Airlines, and RwandAir represent the most advanced NDC implementations on the continent. But advanced by African standards still means early-stage by global benchmarks. The majority of the continent’s carriers remain GDS-dependent, agency-driven, and operationally constrained in ways that make NDC adoption a multiyear undertaking rather than a configuration decision. Travel agencies in most African markets have not resolved the BSP transition problem that NDC introduces. Aggregator coverage for African airline content remains thin. The infrastructure that would allow a coherent African NDC ecosystem to function is still being built.
Against that backdrop, the emergence of MCP as the next interface layer is not a theoretical concern. It is a practical timing problem. If the global industry moves toward AI-agent-driven distribution over the next three to five years, and if that transition is structured around MCP-compatible systems, then African airlines face a version of the same displacement risk that NDC was supposed to solve. Control over distribution, barely recovered from the GDS era through hard-won NDC progress, passes again to whoever controls the orchestration logic. The difference is that instead of Amadeus or Sabre setting the rules, it may be an AI platform operating at a layer African carriers cannot easily influence.
This is not a call to abandon NDC investment. The content and offer management infrastructure that NDC builds remains valuable regardless of how it is eventually surfaced. An airline with well-structured NDC content is better positioned for an MCP-compatible future than one without it. The concern is strategic sequencing, not technology itself.
The question African aviation leadership needs to be asking is not whether to adopt NDC. It is whether the NDC investments being made today are being designed with the next interface layer in mind, and whether the commercial and technical relationships being built now with aggregators, technology partners, and AI platform developers will give African carriers any leverage when orchestration becomes the primary competitive dimension.
The history of airline distribution is a history of infrastructure built by others and tolerated by airlines until the cost of dependence became impossible to ignore. NDC was the industry’s attempt to interrupt that pattern. MCP may be the moment that determines whether the interruption held. For African carriers still mid-transition, NDC risks becoming irrelevant before it ever became useful.
