LO
Travel agent / advisor — tasks, one by one
The unit of analysis is the task, not the job title. Each one below carries its direction, whether the judgement rests on evidence or on platform inference, the reasoning, and what it does not establish.
Every task on this page#
Finding and comparing options
Automating≈ Platform inferenceSearching flights, hotels and packages and comparing them for a client.
This is the part travellers increasingly do themselves. The US statistics bureau says the ability of travellers to use online resources, including AI tools, to research vacations and book their own trips may limit demand for travel agents, and projects little or no change in their employment from 2025 to 2035.
A projection for one country; it names AI as one online resource among others and does not measure how much searching agents still do.
Building complex itineraries
Being augmented✓ Evidence-backedFitting flights, hotels, transfers and budgets into one plan that meets every constraint.
Keeping every constraint at once is exactly where language models fail. In the TravelPlanner benchmark GPT-4 achieved a success rate of 0.6%, and a later paper reports o1-preview at 10% — rising to 93.9% only when the model hands the problem to a formal solver. Agencies are wiring AI in beside people: Flight Centre says AI agents construct the itinerary alongside the consultant, and Trip.com says its AI products support itinerary planning.
Benchmarks in test environments and companies' descriptions of their own tools; neither measures itinerary errors in real bookings.
Booking and ticketing
Automating✓ Evidence-backedMaking reservations, issuing tickets and processing payments.
The transaction is where companies say they are taking people out. Flight Centre told investors it uses AI in key parts of its transactional workflow to remove human touch and lower costs, moving low-value activities to self-service or full AI automation, and reported a leaner corporate workforce alongside system consolidation. Amex GBT says it is implementing automation to increase servicing efficiency.
Companies' own descriptions; the staff reductions they report have several causes, and AI's share cannot be separated.
Changes and disruptions
Being augmented✓ Evidence-backedRebooking after cancellations, handling changes, and helping travellers who are stuck.
Self-service is reaching disruptions, but responsibility and headcount have not gone. American Airlines has tested a generative AI chat assistant that helps customers rebook after weather delays; Expedia says it supports travellers through a virtual agent platform as well as phone and chat. Trip.com reported 16,718 people in its customer service centres, more than a year earlier. Under EU law the organiser of a package is responsible for its performance and must give appropriate assistance to a traveller in difficulty.
An airline test, companies' descriptions and a law that assigns duties to a business, not to a person; none measures how disruptions are handled.
Selling and advising on complex trips
Still human-led✓ Evidence-backedUnderstanding what a client wants and selling trips that are complicated, expensive or for groups.
Where the trip is complex, companies describe AI backing the consultant rather than standing in for them. The US statistics bureau says AI may allow agents to perform administrative and marketing tasks more efficiently. Flight Centre calls its consultants its human superpower, backed by AI, and reports sales per consultant up about 30% since 2023.
A projection and one company's presentation; the productivity figure is the company's own measure and does not show what happens to consultant numbers.