Travel agents

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

• Railway timetables + personal negotiati…1841–1890

  Railway timetables + personal negotiation (Thomas Cook era)

  The first travel agents had no proprietary tools. Thomas Cook's initial advantage was not information, railway timetables were publicly available, but negotiation: the ability to commit to bulk purchase of seats on behalf of a group and extract a discounted fare in return. His circular notes (1865), coupons redeemable at partner hotels across Europe, were the era's closest analog to a booking system: a paper-based guarantee of accommodation that eliminated the uncertainty of arriving at a European hotel without a reservation. The agent's value was entirely interpersonal: relationships with railway managers, hotel keepers, and steamship pursers, plus the ability to translate the complexity of international itineraries into a product a client could purchase without becoming an expert themselves.

  Effect on the work

  Created the occupation from scratch. By 1880 Thomas Cook & Son had offices on multiple continents and had organized the first round-the-world tour. The company was the largest travel intermediary in the world by a substantial margin.

• American Express Traveler's Checks1891–1960

  American Express Traveler's Checks (1891) + ASTA formation (1931)

  J.C. Fargo's frustrated 1888 European trip, in which his letters of credit were routinely refused by foreign banks, led American Express employee Marcellus Berry to design the Traveler's Check, launched in 1891. The product transformed international travel by eliminating one of its central anxieties: arriving in a foreign country without access to locally negotiable currency. For travel agents, American Express Traveler's Checks became a standard product to recommend and sell to international clients, generating both commission income and a trusted supplier relationship. American Express established its first formal travel agency in 1915, entering the intermediary market that Thomas Cook had pioneered. ASTA was founded April 20, 1931 (originally the American Steamship and Tourist Agents Association, reflecting the occupation's still-heavy dependence on Atlantic steamship crossings) and gave the fragmented travel agency industry its first trade association, code of ethics, and lobbying presence.

  Effect on the work

  The 1931-1960 period was dominated by steamship and rail travel; airlines were expensive and rare. The occupation was stable but small, primarily serving wealthy clients who could afford international travel. The democratization of air travel after 1958 (jet service, declining fares) would trigger the occupation's growth phase.

• Manual airline ticket stock1960–1976

  Manual airline ticket stock (ARC/IATA) + airline commission system (10%)

  Before the GDS era, a travel agent's technological apparatus was the Airline Reporting Corporation (ARC) stock of blank ticket blanks (controlled documents, similar to checks) plus a filing cabinet of airline tariff guides (large printed volumes updated monthly), Official Airline Guide issues for schedule lookup, and a telephone. Booking an international itinerary required calling each airline's reservations line, confirming space verbally, then manually writing a multi-coupon ticket using the correct fare constructions from the tariff guide. International airfares were set by IATA and strictly regulated; finding the lowest legal fare required knowing which combinations of fares were permitted. This knowledge barrier was high, and it was the core of the travel agent's value: interpreting the tariff and constructing a valid complex ticket was beyond the capability of most consumers.

  Effect on the work

  The commission model (airlines paid 10% of ticket value to the booking agent, tour operators paid 10-20% for package holidays) meant travel agents had no need to charge clients directly. The consumer got free expert advice; the agent earned from the supplier. This model created extraordinary growth incentives: every new client meant new commissions at no incremental cost.

• GDS terminals1976–1996

  GDS terminals: Sabre (1976), Galileo, Worldspan, Amadeus

  On March 15, 2000, Sabre officially separated from American Airlines, but the critical moment had been 24 years earlier, in 1976, when American Airlines first connected travel agency offices directly to the SABRE reservation system via leased telephone lines and dedicated terminals. The transformation was immediate and dramatic. Before 1976, the average airline reservation took 90 minutes of phone calls; after Sabre terminal access, the same booking took seconds. By the mid-1980s, approximately 50,000 travel agencies across the United States were connected to SABRE alone, and the system processed 45% of all US airline reservations. Competing GDS systems launched: Galileo (United Airlines origin), Worldspan (Delta/Northwest/TWA), and Amadeus (Air France/Iberia/Lufthansa/SAS). The GDS era democratized access to professional travel agency tools: any agency, in any city, with a terminal and a lease, could access global airline inventory. The economics were extraordinary: airlines paid not just commissions but also booking fees to the GDS, which passed a portion to travel agencies in the form of preferred-supplier incentives. At peak, in the early 1990s, a profitable US travel agency's revenue came primarily from airline commission (10% of ticket value) plus override incentives from preferred suppliers for hitting volume targets, plus GDS productivity bonuses. The entire revenue model required zero direct billing of clients.

  Effect on the work

  Employment grew through the GDS era. The GDS terminal made small independent agencies economically viable for the first time; the US travel agency count grew from a few thousand in the pre-GDS era to over 30,000 agencies by 1990.

• Internet OTAs1996–2010

  Internet OTAs: Expedia (October 1996), Travelocity (1996), Priceline (1998), Orbitz (2001), Kayak (2004)

  Expedia launched as a Microsoft division in October 1996: the first large online travel agency, offering consumers direct access to the same airline inventory that travel agents had accessed via GDS terminals. Travelocity, built on the Sabre GDS by Sabre Corporation, launched the same year; the Wikipedia entry on Travelocity describes it as "the first website that allowed consumers the ability to purchase travel tickets without the help of a person." Priceline launched in 1998 with its "Name Your Own Price" opaque-inventory model. Orbitz launched in 2001, backed by the airlines themselves. Kayak launched in 2004 as a metasearch aggregator. The internet OTAs did not just provide convenience: they eliminated the information asymmetry that had been the travel agent's core value proposition for GDS-accessible, commoditized bookings. Airlines simultaneously cut and then eliminated agent commissions: Delta led the cuts in 1995, and by 2002 virtually all US domestic airlines had reduced commissions to zero for standard tickets. The combination destroyed the economic foundation of the GDS-era travel agency model. BLS counted 111,130 US travel agents in its OES 1999 national estimates and 87,600 in May 2006.

  Effect on the work

  Employment fell between the 1999 and 2006 BLS counts. A larger number of small independent agencies closed entirely; the surviving workforce concentrated in corporate travel management companies (where employer relationships and duty-of-care obligations kept agents in the loop) and high-end leisure specialists who could charge fees for complexity.

• Mobile booking apps + Airbnb2010–2023

  Mobile booking apps + Airbnb (2008) + stabilized niche positioning

  Airbnb launched in 2008 and by 2015 had listings in 34,000 cities, completing the disintermediation of accommodation bookings that hotels.com and booking.com had begun for traditional hotels. Mobile booking apps (Hopper, Google Flights, TripAdvisor) further commoditized price comparison. Yet travel agent employment stabilized in this period: approximately 70,000-85,000 agents persisted through the 2010s, an occupational floor that the continued launch of new disintermediation tools failed to breach. The stabilization is explained by segment bifurcation. Cruise lines maintained travel agent distribution as a primary channel: CLIA (Cruise Lines International Association, founded 1975) enrolled 13,000+ travel agencies and 50,000+ individual travel agent members, and cruise lines historically sourced the majority of their bookings through agents because complex cabin-class decisions and shore-excursion coordination benefited from personal guidance. Corporate travel management remained agent-mediated because employers needed consolidated billing, duty-of-care tracking, policy enforcement, and managed vendor relationships: services that consumer OTAs did not offer. The Virtuoso luxury travel network (which describes itself as a network of luxury travel agencies) provided high-end agents with preferred access to hotel amenities, suite upgrades, and exclusive experiences that OTAs could not replicate. The surviving agents became specialists rather than generalists.

  Effect on the work

  Employment held at ~70,000-85,000 through 2018. The occupation's compositional center of gravity shifted decisively toward complex, high-value bookings: cruise itineraries, luxury resort stays, safari packages, round-the-world itineraries, corporate account management. The transactional booking segment was effectively gone.

• COVID-192020–2022

  COVID-19: full industry shutdown (March 2020)

  In March 2020, the US State Department issued a Level 4 "Do Not Travel" advisory: the first blanket global travel warning in the department's history. Airlines cancelled 95%+ of international flights. Cruise lines shut down entirely under CDC No Sail orders. Hotels closed. Travel agents experienced an immediate and total cessation of new bookings accompanied by a wave of cancellations: refund and rebooking work at zero commission. BLS OEWS recorded 55,180 travel agents in May 2020, a 17% drop from the 2019 level. The scale of layoffs, furloughs, and agency closures exceeded anything the internet had produced in a comparable period. Many experienced agents left the occupation permanently; many small independent agencies closed without reopening. Those who survived typically had either strong corporate accounts (which continued managing stranded-traveler situations, repatriation logistics, and eventual recovery bookings) or strong luxury client relationships (whose clients intended to travel again and valued continuity with an agent who knew their preferences).

  Effect on the work

  The single largest employment disruption in the occupation's modern history. The 2020 employment level (55,180) was below the pre-internet GDS era workforce. Recovery was partial: by 2024, employment had recovered to 59,150, approximately 89% of the 2019 level.

• AI itinerary tools2023–present

  AI itinerary tools: ChatGPT travel planning, Mindtrip, Expedia GenAI chatbot, Layla AI

  The second wave of travel agent disintermediation began in 2023, with a different character than the first. The 1996-2006 wave eliminated information asymmetry around standardized products (airline tickets, hotel rooms). The 2023+ wave targets the complex itinerary and relationship layers that survived the first wave. Mindtrip (an AI travel planning platform backed by $7 million in funding, partnered with Priceline, TripAdvisor, Google Places, and Viator) offers conversational itinerary generation, multi-leg planning, and booking integration. Expedia launched an AI-powered chatbot in 2023 integrated with ChatGPT. The structural question is whether AI tools can replicate the travel agent's value in relationship segments (knowing that a particular client always requires a lower-deck cabin because of seasickness; knowing that a corporate client's CEO refuses to fly anything but direct business class; knowing that a luxury hotel's concierge will arrange the private safari access that is not listed anywhere). Current AI tools are effective at generating first-draft itineraries for destination-agnostic leisure travel; they are weak at managing the supplier relationships, accountability during disruptions, and client-preference memory that surviving agents provide. The second disintermediation attempt is real, but the target segments are smaller and more resilient than what the internet found in 1996.

  Effect on the work

  Too early to measure in BLS employment data as of 2024-2026. The BLS 2024-34 projection of +2.2% employment growth suggests the agency does not model AI tools as net-negative for travel agent employment at current trajectory, consistent with the view that the remaining occupation is concentrated in segments where AI tools assist rather than substitute.