Travel Agency Stocks In An AI World: Investors Think There’s No Future, But Is The Panic Unnecessary?

Travel agency companies survived two world wars, the jet age, the internet, and COVID-19. It may yet survive artificial intelligence, but clearly the market is not waiting around to find out. Just look at the share price declines of companies like Flight Centre (ASX: FLT) and Web Travel Group (ASX: WEB) on the ASX, not to mention companies overseas such as Booking Holdings (NASDAQ: BKNG).

Just about all have absorbed meaningful losses over the past year as investors grapple with an increasingly coherent threat to the intermediary model that underpins the travel distribution industry.

The sell-off in these companies is real, even if the ultimate outcome remains genuinely uncertain.

Why the market is nervous about travel agency stocks

The AI threat to online travel agencies (OTAs) and traditional travel agents is not purely hypothetical. It rests on a clear and commercially logical mechanism: if an AI agent can understand a traveller’s preferences, search live inventory, compare prices, and execute a booking (all within a single conversational interface) then the role of the intermediary is structurally diminished. The platform that owns that interface captures the economic value that previously accrued to Booking.com’s search rankings, Flight Centre’s storefront, or Web Travel’s wholesale hotel beds.

This casts our minds back to Citrini Research’s dire warning about AI that rippled markets in February, modelling a future scenario in mid-2028. It warned any business model where the add on is ‘we will help you navigate complexity’ is doomed because AI agents find nothing complex. Travel agents are not the only business facing this risk, but they are facing the risk.

And we’re not talking hypotheticals – that shift is underway, the AI giants are coming after their prey. OpenAI spent much of 2025 building native booking capabilities within ChatGPT, including a “buy now” function aimed squarely at travel, before pulling the feature in early 2026 after discovering that users were reluctant to commit credit card details to an unfamiliar checkout environment. Google, similarly, has been developing agentic AI tools for flight and hotel search. It has publicly stated that it has no intention of becoming an OTA itself, but it is a reassurance that did little to stop the sector re-rating downward when the news broke.

By some industry estimates, more than half of travellers were already using AI tools such as ChatGPT at some stage of trip planning by the end of 2025. And 6 months into 2026, it wouldn’t be unreasonable to assume even more would be. But even if it was just over half (specifically 54% was estimated), that is a meaningful share of the discovery and research phase that OTAs have traditionally owned.

The story behind individual stocks

Flight Centre has had an especially difficult eighteen months. The company’s shares have shed approximately 35% over the past twelve months, with the stock trading near A$10.60 at the time of writing against an analyst consensus target of A$16.63, representing a gap that suggests either the sell-off is overdone or the earnings trajectory has structurally deteriorated.

In FY25, FLT’s underlying profit before tax came in at A$289m, a decline of roughly 10% on the prior year and well below guidance that had originally pointed toward A$365–405m. The company also guided for flat underlying PBT in 1H26 compared to market expectations of 6% growth, prompting Jarden analysts to flag “mid-single digit 1H26 consensus cuts.”

Macro headwinds, including US tariff uncertainty and softer consumer confidence, bear much of the immediate blame. But at the company’s AGM in late 2025, chairman Gary Smith was unusually candid about the technology dimension, acknowledging that while AI presents Flight Centre with opportunity, it “does, of course, also represent a threat as competitive customer offerings emerge.” The company, he said, was “working diligently to combat these competitive threats and offerings globally as they arise.” That is an honest framing; it is not, however, a detailed strategic response.

Web Travel Group tells a somewhat different story, partly because its most acute problems have been company-specific rather than sector-wide. After spinning off its consumer-facing OTA business (now Webjet Group, ASX: WJL) in late 2024, Web Travel retained WebBeds — a B2B hotel distribution platform.

The share price, however, has been in sustained decline, underperforming the ASX All Ordinaries Index by nearly 46% over the past six months and trading around A$2.61 against a consensus target of A$4.49. Margin erosion from the FTI Group collapse in Europe, subdued post-Olympics European trading, and customer override agreements have compressed WebBeds’ TTV/revenue margin. Meanwhile, the separately listed Webjet Group has seen its own booking volumes fall 8% in the most recent half as domestic flight bookings softened, with its share price collapsing 25% in early 2026 after both a failed takeover process and a guidance downgrade.

Even the global bellwether is not immune

Booking Holdings (NDQ:BKNG), the world’s largest OTA, has fallen approximately 23% year-to-date in 2026. Some analysts have linked it specifically to the Anthropic Claude plugin launch — a development seen as a credible direct-booking threat to OTA intermediaries. That decline is notable given that Booking’s operational performance has remained broadly sound: in Q4 of calendar 2025, room nights grew 9%, gross bookings and revenue rose 11% on a constant-currency basis to US$6.3bn, and adjusted EBITDA expanded 19% to US$2.2bn. The stock now trades at approximately 15x forward earnings, which, as one commentator noted, represents an arguably cheap multiple for a business generating close to US$10bn in free cash flow annually.

The valuation compression, in other words, is predominantly sentiment-driven rather than earnings-driven, which is characteristic of a market pricing in a structural risk before it has fully materialised.

What the companies claim to be doing about it

The corporate responses range from substantive to reassuring-sounding. No CEO will come out and say either they’re not trying or there’s no threat at all. But the most assuring PR narrative won’t soothe investors if it is not backed up by action, and preferably action paying dividends.

Booking Holdings has been the most operationally specific. The company has deployed what it describes as a disciplined, modular AI architecture: small, travel-specific models for fast and cheap inference, larger LLMs for reasoning and contextual understanding, and domain-tuned evaluations to maintain accuracy. Its customer-facing AI products include Smart Messenger, Auto-Reply, AI Trip Support, AI Voice Support, and AI Flight Search Summaries.

The company has also moved to make Booking.com an integrated partner within OpenAI’s ChatGPT app ecosystem, accepting that AI assistants are a new distribution channel rather than purely a competitive threat. The proxy statement that the company filed in 2026 articulates this explicitly: “We envision a seamless and personalized experience in which AI-powered agents help coordinate trips with greater customisation, context, and convenience.” Whether that vision translates into retained intermediary economics, or simply a more frictionless route around Booking’s platform, remains to be tested.

Flight Centre’s response has leaned on a more human-centric argument. The company has highlighted that average consultant revenue has reached A$165,000 and is targeting A$200,000, arguing that AI has amplified rather than replaced agent productivity. The investment in AI innovation announced alongside the FY25 downgrade was described as aimed at “enhancing customer experience, boosting productivity and disrupting traditional offerings.” That framing is strategically sensible: the argument that an AI-augmented human travel consultant provides more value than a raw AI agent is plausible, particularly in corporate and complex travel segments where the corporate division reported a record A$6.3bn in total transaction value and 20% divisional profit growth in 1H26.

Web Travel’s position is structurally more complex. A B2B hotel distributor faces a somewhat different AI exposure than a consumer OTA: the risk is less that a consumer chatbot bypasses WebBeds directly, and more that AI-enabled procurement tools allow hotel chains and accommodation providers to reduce their reliance on wholesale intermediaries altogether. Management has not disclosed a detailed AI strategy specific to WebBeds, and the company’s near-term investor narrative remains dominated by margin recovery rather than technology positioning.

Separating noise from signal

In our view, the market has conflated several distinct risks into a single de-rating event. The immediate earnings pressure across all three companies reflects macro headwinds, consumer confidence softness, and company-specific operational issues more than demonstrated AI revenue displacement. Booking Holdings’ financial results, in particular, suggest the disruption narrative is running well ahead of any observable fundamental deterioration.

The longer-term risk is more credible but also more gradual than current share price movements imply. The evidence from OpenAI’s retreat on native checkout suggests that trust, brand, and transaction security remain meaningful moats. AI agents appear well-suited to trip discovery and planning; they are less immediately compelling as transaction executors for high-value, anxiety-laden purchases like international flights and hotels. Booking’s observation that hotel and accommodation data is deeply complex (requiring supplier relationships, content quality, and payment infrastructure that cannot be easily replicated) is analytically sound.

None of that makes the risk zero. It does suggest, however, that investors treating current valuations as pricing a worst-case scenario may be looking at asymmetric upside, provided management teams are building the technical capability to participate in, rather than be bypassed by, the agentic AI distribution channel. On that score, Booking Holdings appears furthest advanced. Flight Centre is making the right arguments, even if the execution evidence is thinner. Web Travel, for now, has other problems to solve first.