A global study on the rise of AI agents in travel
The report “Win the AI Agent, Win the Traveler” draws on two separate surveys conducted in June 2026. The first gathered responses from 5,003 travellers across twelve markets: the United States, Canada, the United Kingdom, Germany, France, Spain, Italy, India, China, Japan, the United Arab Emirates and Saudi Arabia. The second surveyed 600 executives from airlines, hotel groups and online travel agencies, based in seventeen countries across Asia-Pacific, Europe, North America, the Middle East and Latin America.
Sixty per cent of these executives represent companies with annual revenue above €870 million (around $1 billion). The report also draws on two earlier Accenture studies, Agentic Commerce: Make Your Brand Unmissable and Consumer Pulse 2026, as well as on the firm’s client engagements in the sector.
Travellers ready to delegate, with conditions
The complexity of the booking journey remains the primary friction point identified by Accenture. Nearly two in three travellers use at least four different platforms to organise a single trip, and 94% consider the process moderately or very complex. Two-thirds believe this complexity has caused them to miss better options.
This friction is turning simplicity into a genuine deciding factor: 96% of travellers surveyed say they are willing to switch away from a habitual travel brand if an alternative offers a noticeably faster, simpler experience. When it comes to AI agents, however, delegation remains measured and proportionate to perceived risk. Forty per cent of travellers are willing to hand comparison-shopping entirely to an agent, versus 29% for full trip management, 24% for loyalty optimisation, 14% for finalising a booking, and just 10% for executing payment.
Loyalty follows the same logic of flexibility. Eighty-three per cent of travellers want to be able to spend their rewards beyond the travel sector alone, and 86% would be more inclined to choose a brand whose benefits combine with their payment methods or other loyalty programmes. Crucially, 96% value the idea of a single system bringing together payments, rewards and preferences, and 97% would be willing to use it even if that meant sharing personal and payment data.
An internal architecture that remains fragmented
Faced with this demand, the industry is lagging badly on execution. Only 30% of executives surveyed consider their organisation fully ready to orchestrate an end-to-end customer journey across partners and ecosystems. The systems AI agents need to be able to draw on remain largely siloed: customer data is seen as fragmented by 87% of executives, payments by 86%, identity and profiles by 79%, and loyalty by 72%.
Executives point to three main obstacles to connecting these systems: organisational complexity (70%), legacy systems (65%) and data silos (65%). Yet AI investment is not following this priority: 66% of companies are already deploying AI in marketing, compared with just 23% in payment systems — despite the latter being identified as the most critical point of failure for a transactional agent.
This fragmentation carries a measurable cost. More than eight in ten executives believe it has a significant impact on revenue growth (88%), on control of the customer relationship and on conversion rates (84% each), and on operating costs (83%). The financial scale of loyalty illustrates the sums at stake: the remuneration paid to Delta Air Lines by American Express under their co-brand partnership reached $8.2 billion in 2025 — roughly €7.16 billion — up 11% year-on-year.
🔗 More on LTH: our analysis of hotel distribution and where OTAs stand as AI reshapes the sector
The risk of losing the customer relationship to agents
OpenAI’s retreat from its direct payment option within ChatGPT illustrates just how hard it is to execute a reliable end-to-end transaction. The company itself acknowledged that the feature did not offer the level of flexibility expected, and redirected transactions to third-party applications. For travel — where fare volatility, cancellation rules, multi-currency payments and post-booking service are added into the mix — the technical bar is even higher.
Traveller demand, however, goes beyond simply wanting a more pleasant interface: 71% would like an AI agent to coordinate an entire trip, including flights, hotels, car hire, itineraries and logistics. Meeting that expectation requires what Accenture calls “value orchestration”: the ability to move information between systems without artificially merging them.
The reliability stakes are immediate. According to an Accenture simulation involving 50,000 consumer agents, 86% of agent-mediated transactions risk being abandoned or redirected to a competitor if something goes wrong at the point of purchase, with travel among the most exposed categories. This has already happened in practice: in 2025, ChatGPT’s agent mode carried out an end-to-end booking with a US online travel agency without the user ever leaving the chat interface. The transaction itself was executed by the OTA, but it was ChatGPT that kept the relationship with the traveller.
Three priorities, depending on the link in the value chain
For airlines, the priority is to bring the traveller’s profile, preferences, entitlements, loyalty status and transaction history together on a single foundation across every touchpoint. This foundation is what enables the subsequent modernisation of payments and loyalty, as well as better handling of disruptions. Air India illustrates the approach: its 2026 app overhaul brought booking and payments back in-house, with generative AI deployed across the entire journey rather than outsourced to third-party providers.
For hotels, the urgent task is to make their value legible to a machine. Rates, member conditions, upgrade rules, points-conversion value and cancellation policies are still too often scattered across systems designed for humans rather than for agents. Singapore’s Ascott Group has been restructuring its systems to surface its inventory directly within major AI platforms, allowing travellers to discover, compare and book its properties through the AI agent of their choice.
For OTAs, which already aggregate supply, the path to an agent-compatible model may be shorter, but the competitive risk is higher: if a generative AI platform becomes the traveller’s point of entry, the OTA risks being reduced to handling the technical side of the transaction, while the customer relationship, the data and the upsell opportunities shift to the layer above. Accenture recommends starting with payment orchestration, built on a reliable multi-currency, multi-method layer capable of surfacing the full range of options to agents.
Reliable execution: the new competitive advantage
Connectivity alone will not be enough. Nearly four in five executives cite security risks and regulatory compliance among their top concerns, alongside governance, data reliability and customer trust, all seen as essential before allowing AI to take action on a customer’s behalf. Even so, executives expect the shift to come quickly: around three in four believe AI will be carrying out concrete customer-facing actions — including bookings, payments and disruption management — within three years.
For Accenture, reliable transaction execution is on its way to becoming as decisive a differentiator as brand, loyalty or price. Companies that delay connecting their systems may still survive within the travel ecosystem, but on terms set by others.
At a glance by The Hospitality Tribune
Accenture Research: travel companies race to become agent-ready
Survey base: 5,003 travelers across 12 markets and 600 travel executives across 17 countries (June 2026)
Only 30% of executives feel fully ready to orchestrate end-to-end journeys across partners
Core systems remain siloed: customer data (87%), payments (86%), identity (79%), loyalty (72%)
71% of travelers want an AI agent to coordinate a full trip; 96% would switch brands for an easier experience
86% of agent-mediated transactions risk abandonment when a purchase goes wrong, per Accenture’s Aaru simulation
Delta’s American Express remuneration reached $8.2 billion in 2025, up 11% year-over-year
