The obituary has been written many times. First it was the internet. Then it was the OTA. Then it was COVID, which was supposed to finish the job. The travel agent survived all of it. Now the industry is back to writing the same obituary, this time with artificial intelligence as the cause of death.
The agents actually paying attention are not reading the obituary. They are reading the tools.
There is a version of the AI disruption story that is easy to tell. Autonomous booking agents. Conversational trip planners that never sleep. Systems that can search 200 fare combinations in the time it takes a human agent to open a GDS terminal. On paper, the math looks bad for the human in the middle.
But that version of the story makes a mistake that the industry keeps making about travel agents: it assumes they are passive. It assumes the disruption arrives and the agent simply absorbs it.
That has not been the pattern. It was not the pattern with OTAs, which were supposed to eliminate agents entirely and instead created a market bifurcation where complex, high-value itineraries kept flowing through human hands. It is not the pattern now.
The TMCs that are scaling in 2026 are not scaling despite AI. They are scaling because of how they have positioned it.
FCM Travel is rolling out a fundamentally reimagined version of its AI tool Sam, designed to work across travelers, travel managers, and agents simultaneously, not as a replacement layer but as an operational one. The distinction matters. Sam does not replace the agent relationship. It removes the friction that was consuming the agent’s time before the relationship could even begin.
That is the operative insight. The agents losing ground to AI are the ones who treated it as competition. The ones gaining ground treated it as the thing that finally gives them capacity to do what they were always better at: judgment, relationships, and handling the situations that no algorithm has been trained to resolve at 11pm when a client is stranded in Lagos and the rebooking queue is three hours long.
The data is worth sitting with for a moment.
According to the Airlines Reporting Corporation, OTAs account for 77% of settled NDC transactions. Corporate agencies account for 7%. Leisure agencies account for 16%. The gap between OTAs and traditional agents on NDC is not primarily a technology gap. It is a capacity gap. OTAs had engineering teams. Agents had GDS terminals and a BSP account.
The significance is not that AI creates NDC adoption automatically. It reduces the operational burden that has kept smaller agencies on the sidelines — client communications, offer comparison, post-booking servicing — so that NDC participation stops requiring an engineering team. The agency that could never justify hiring a developer to build an NDC integration can now close much of that gap through tools that did not exist at an accessible price point three years ago.
AI changes that equation in ways that are still being underestimated. The same tools that allow a large OTA to run automated search and booking pipelines are now accessible to a three-person agency in Nairobi at a subscription price that did not exist three years ago. The infrastructure cost that separated the two has compressed dramatically.
The agent who understands this is not looking at AI as a threat to their existence. They are looking at it as the first technology in a decade that actually tilts the playing field back in their direction.
There is an Africa-specific dimension here that the global conversation consistently misses.
The framing in European and North American markets is almost always about legacy agencies with deep GDS infrastructure, long-standing corporate contracts, and established workflow dependencies. Disrupting that model is genuinely expensive and slow because the cost of switching is high on both sides.
African agencies largely do not have that problem. Most are running leaner operations, with lower infrastructure lock-in and a client base that expects their agent to be responsive across WhatsApp as much as across any formal booking system. The barrier to adopting AI-assisted tools is lower precisely because the prior investment in legacy infrastructure was never as deep.
That is not a consolation prize. It is a structural advantage that the market has not priced correctly.
An agency in Accra or Kampala that integrates AI-assisted itinerary building, automated client communication, and NDC content access through a modern aggregator is not catching up to the global market. In several meaningful respects, it is ahead of a mid-size European agency still running EDIFACT workflows and waiting for its GDS provider to finish a migration it started two years ago.
The agents who will struggle over the next three years are not the ones competing with AI. They are the ones who refused to engage with it until the choice was no longer theirs to make.
The ones who will win are already asking a different set of questions. Not “will AI replace me” but “which parts of my job should AI be doing so I can focus on the parts it cannot.”
That reframe is available to any agent who wants it. The technology is not the barrier. The mental model is.
The travel agent is not dying. The version of the travel agent who treated every technology shift as someone else’s problem is. The version who moved first is building something the algorithm cannot replicate: trust, earned over years, with clients who know the difference between a system that booked their flight and a person who got them home.
