Ticket and reservation agents

Three parts of your work where AI is already doing real lifting, and what stays yours.

• Telegraph + paper ticket stock1840–1930

  Telegraph + paper ticket stock (railroad station agent era)

  The railroad station agent operated the telegraph to receive train orders and schedule changes, and sold paper tickets from a standardized stock printed by the railroad company. Reservations for sleeping car berths were coordinated by wire between station agents. There was no centralized seat inventory: an agent selling a berth in a Pullman car had to wire ahead to confirm availability, often with overnight turnaround. The combination of telegraphy and paper tickets was the first technology system that made it possible to sell a seat on a vehicle before the customer was physically at the departure platform.

• Manual reservation card-file systems1930–1964

  Manual reservation card-file systems (airline and hotel pre-computer era)

  As commercial aviation grew from the late 1920s onward, airlines invented a manual card-file reservation system: each flight was represented by a rotating card drum or board showing available seats. Reservation agents at telephone desks would pull a card, mark a seat, and relay the information by phone to other offices. American Airlines' "Reservisor" system of the early 1950s was a semi-mechanical version that stored seat availability on a central drum accessed by phone. The process of confirming a seat still took minutes end-to-end, required multiple telephone calls between offices, and was highly error-prone. The reservation agent in this era was primarily a telephone operator who navigated a complex manual information system.

  Effect on the work

  The card-file era required large reservation centers staffed with dozens of agents per airline per city. American Airlines alone maintained thousands of reservations agents nationwide by the late 1950s, all working phone banks with manual systems.

• SABRE1964–1980

  SABRE (Semi-Automated Business Research Environment): first real-time CRS

  SABRE went live in 1964 as the world's first fully operational computer reservations system, developed jointly by American Airlines and IBM after a chance meeting between American's president C. R. Smith and an IBM salesman on a flight in the mid-1950s. Two IBM 7090 computers at Briarcliff Manor, New York handled real-time processing and batch backup. Agents accessed the system through Raytheon terminals. The immediate effect was transformative: average reservation time dropped from many minutes to a few seconds. SABRE could handle thousands of reservations per hour. For the reservation agent, the CRS terminal replaced the manual card system and the relay telephone call, creating a new skill requirement: proficiency on a specialized computer terminal with cryptic command codes. The role became more technical, more urban-office-concentrated, and more productive per agent.

  Effect on the work

  SABRE reduced the labor required per reservation transaction substantially, but demand for air travel grew faster than productivity gains through the 1960s and 1970s, so overall reservation agent employment continued to increase during the SABRE era rather than contracting. The system enabled the airline industry's expansion rather than eliminating the workforce that supported it.

• GDS distribution to travel agencies1976–1996

  GDS distribution to travel agencies (Sabre 1976, Apollo/Galileo, Amadeus)

  In 1976, American Airlines began offering SABRE access to outside travel agents for the first time, transforming the CRS into a global distribution system (GDS). Competing systems followed: United's Apollo (1971, opened to agents 1976), PARS (later Amadeus), and Worldspan. For the first time, travel agents outside airline offices could see and book seat inventory directly. This created a new workforce category: the travel agency booking clerk who used GDS terminals daily. For airline counter and phone agents, the GDS era meant that much routine booking work was pre-handled by travel agencies; counter agents shifted toward face-to-face check-in, ticket exchange, and upgrades. The GDS terminal's cryptic command language (e.g., "1Y10JULSFOORD/Y" to book a coach seat) became the defining skill of both agency and airline reservation staff.

• Consumer web booking1996–2010

  Consumer web booking (Travelocity March 1996, Expedia October 1996)

  On March 12, 1996, Sabre launched Travelocity.com as the first website allowing consumers to purchase airline tickets directly without a human agent. Microsoft launched Expedia on the Microsoft Network in October 1996. These two launches began the most consequential structural shift the occupation had experienced since SABRE: the transfer of the standard ticket transaction from a human counter or phone agent to the consumer themselves. The American Society of Travel Agents initially dismissed the threat, issuing a statement that "most people think their time is too valuable" to book online. They were wrong. Within a decade, the majority of US airline tickets were booked online. The resulting displacement was severe: airline counter and phone reservation agent employment peaked around 2000 and then contracted sharply as online self-service absorbed the routine domestic booking transaction.

• Self-service kiosks2008–2020

  Self-service kiosks and mobile check-in (IATA CUSS 2004; airline apps 2010s)

  IATA's Common Use Self-Service (CUSS) kiosk standard, adopted from 2004 onward, enabled airports to deploy shared self-service check-in terminals usable by multiple airlines. Airline mobile apps with mobile boarding passes followed from around 2010, allowing passengers to check in on their phones a day before departure without any agent interaction. By the mid-2010s, most airline check-in transactions in major US airports were handled by self-service kiosk or mobile app. The net effect on the gate agent workforce was a shift in job content: agents who had spent most of their shift processing standard check-ins now spent the majority of their time handling the exceptions that kiosks could not: identity document issues, oversized baggage, unaccompanied minors, and accessibility needs. The volume of face-to-face transactions dropped substantially even as the complexity of each remaining transaction increased.

• COVID-19 disruption2020–2022

  COVID-19 disruption: furloughs, workforce reduction, remote reservation migration

  In March and April 2020, US airline passenger volumes collapsed from year-earlier levels. American Airlines furloughed 19,000 workers and United furloughed over 13,000 in October 2020 alone after the CARES Act payroll support expired. Passenger-handling occupations including ticket agents and gate agents comprised a large share of the total major airline workforce and were identified in research as "the first line of cost shedding" during the crisis. November 2020 was the lowest airline employment month on record since tracking began in 1990. The furloughs accelerated several structural changes: airlines moved more reservation functions to remote work or chatbot handling, reduced the number of staffed counters at airports permanently, and emerged from the pandemic with leaner agent workforces.

  Effect on the work

  Tens of thousands of aviation workers including ticket and gate agents lost their jobs or were furloughed in the fall of 2020. BLS OEWS showed a sharp decline in 43-4181 employment in 2020-2021 relative to 2019, with only partial recovery by 2024.

• AI rebooking assistants2023–present

  AI rebooking assistants and agentic travel booking (Sabre Mosaic, Amadeus Cytric AI, Google agentic travel)

  Sabre Mosaic Automation Manager, Amadeus Cytric AI Assistant for Microsoft Teams, and agentic travel tools from Google, Expedia, and Booking.com entered commercial deployment from 2023-2025. These tools automate the queue-based ticketing work that reservation agents previously handled manually: exchange and refund processing, proactive disruption rebooking, and corporate travel policy enforcement. Sabre claims up to 72% efficiency gains on repetitive ticketing queue tasks. Google announced plans in 2025 to build a fully agentic travel booking system capable of completing end-to-end trip reservations autonomously. IDC projected that 30% of travel bookings could be handled by AI agents by 2030. For the remaining human reservation agent, the value proposition is increasingly concentrated in the cases AI cannot handle: complex multi-carrier international itineraries, passenger emotions during a missed connection, and the physical judgment of a gate hold.