Travellers in the United Arab Emirates and Saudi Arabia are increasingly comfortable letting artificial intelligence pick and package parts of their trips: a new Marriott Bonvoy study finds roughly two‑thirds would
trust AI to book accommodation, and nearly three‑quarters have already used AI to research or plan travel — signals that generative tools have moved beyond novelty into mainstream travel planning.
Background / Overview
Marriott Bonvoy’s 2026 Ticket to Travel research surveyed more than 4,000 adults across the UAE and Saudi Arabia and paints a picture of resilient demand for holidays next year. The study reports that around 84% of UAE respondents and 83% of Saudi respondents expect to take
more or the same number of holidays in 2026 compared with 2025, and over half anticipate travelling more frequently. Respondents plan an average of five trips in 2026 (a mix of domestic, short‑haul and long‑haul), and commonly book about two months ahead. The report highlights two consumer trends flagged for 2026:
lux‑scaping (adding a high‑end stay or spa at the start or end of a trip) and
passion pursuits (holidays organised around personal interests such as sports, music, or adventure). It also notes platform preferences for AI planning: ChatGPT tops the list, followed by Google Gemini and Microsoft Copilot. These takeaways are consistent with Marriott’s regional EMEA release and coverage from trade outlets.
What the numbers show: adoption, trust and booking habits
- 72% of UAE and Saudi respondents said they have used AI to research or plan a trip; 23% say they use AI “all the time.” ChatGPT is the most used platform.
- Around two‑thirds (roughly 67–71% depending on market breakdown) say they would feel comfortable booking accommodation through AI platforms; only about 9% reported being uncomfortable with that idea.
- The typical Gulf traveller in the survey plans three domestic trips, two short‑haul breaks, and two long‑haul holidays in 2026, with Italy, Switzerland, France and Türkiye among the most popular international choices.
These are not isolated data points: Marriott’s broader EMEA research and region‑specific press materials show parallel growth in AI usage for travel planning, underlining a genuine behavioural shift rather than a media spike.
Why AI is sticking in travel planning
AI’s rapid adoption for travel tasks is explainable and pragmatic: it accelerates ideation, personalises suggestions, and cuts the hours of browsing that traditional trip research demands. For time‑poor consumers — families, business travellers with leisure add‑ons, and younger cohorts — an assistant that suggests destinations, drafts itineraries, and composes packing lists is a real productivity win.
- Benefits noted by travellers and industry testing:
- Faster destination discovery and themed suggestions.
- Rapid day‑by‑day draft itineraries and packing/checklists.
- Multi‑platform convenience (exporting itineraries to calendars, spreadsheets, or messaging apps).
For travel brands, AI offers scale: personalised marketing, automated pre‑trip communications, and the promise of “agentic” assistants that can take scoped actions on users’ behalf (e.g., draft a booking or propose alternative hotels in real time). This convergence of discovery + action is what many vendors and OTAs are building toward.
Lux‑scaping and passion pursuits: what they mean for travel demand
The Marriott research flags two distinct demand drivers that mix aspiration with experience economy dynamics:
- Lux‑scaping — travellers deliberately tack on a luxury stay or spa session at the start or end of a trip to bookend the experience. More than 80% of UAE and Saudi respondents say they’ve tried it, and many cite relaxation and rejuvenation as the payoff.
- Passion pursuits — trips designed around hobbies or interests (sporting events, music festivals, adventure activities). Over 80% reported having taken such interest‑driven holidays previously; sports travel is notably strong.
These trends matter because they drive
higher‑margin behaviours: travellers are prepared to pay a premium for targeted, high‑value experiences (luxury spas, premium event tickets, specialist adventure guides) — and AI can both inspire and operationalise those purchases.
Strengths: where AI genuinely adds value to travellers and the industry
- Speed and inspiration — AI quickly reduces the “blank page” problem, suggesting destinations, itineraries and themed trips that can spark decisions.
- Personalisation at scale — models can combine expressed preferences (family, budget, activities) with broad destination knowledge to propose tailored options.
- Administrative automation — packing lists, visa reminders, calendar integration and pre‑trip checklists remove friction and reduce no‑shows or misunderstandings.
For hotels and hospitality brands, AI can increase conversion by delivering contextual upsells (lux‑scaping options, spa packages linked to a booking) at the moment of decision — an obvious business upside if executed with correct provenance and pricing.
Risks and failure modes: why comfort ≠ infallibility
The rise in trust is significant, but it must be qualified by the
known technical limits and operational hazards of generative models.
Hallucinations and incorrect operational detail
Large language models are powerful pattern matchers, not guaranteed truth engines. They sometimes produce plausible but false information — invented attractions, incorrect transport schedules, or logistically impossible itineraries. Reported real‑world incidents include travellers being directed to non‑existent sites or arriving after last‑service cutoffs because an AI supplied an outdated timetable. These are not theoretical risks; multiple industry analyses document such failure modes and urge human verification.
Data privacy, training and contractual exposure
Many consumer chat services retain and use prompt data to improve models unless you use a paid or enterprise tier with specific non‑training clauses. Travellers who paste booking references, passport numbers or payment details into generic chatbots may inadvertently expose PII to model training or to insecure retention policies. Corporate travel buyers and privacy officers should insist on contractual guarantees and data non‑training clauses when selecting vendor tools.
Liability, cancellations and accountability
If an AI both recommends and
books travel, where does legal responsibility sit when something goes wrong? The operational lines between suggestion and actuation blur as “agentic” assistants gain permissions to complete transactions. Travel companies, regulators and insurers will need clearer rules about audit trails, consent, and remediation paths when bookings are initiated through automated agents.
Grounding and provenance shortfalls
Reliable travel AI products require retrieval‑augmented generation (RAG) and live connectors to authoritative data sources (airline APIs, official timetables, hotel availability feeds). Without that grounding, the chance of an error that affects safety rises. Industry guidance emphasises visible provenance (source links, “last verified” timestamps) and human checkpoints for safety‑critical outputs.
Cross‑checking Marriott’s headline claims
Marriott’s headline numbers — AI adoption at scale, trust levels and trend signals like lux‑scaping — are corroborated across trade and regional outlets. Arabian Business summarised the Marriott Bonvoy findings for UAE and Saudi travellers; PR and trade coverage of Marriott’s EMEA Ticket to Travel research presents consistent regional patterns (growing AI adoption, younger cohorts leading the shift, and increased demand for experiential add‑ons). Where small differences exist between outlets, they typically reflect sample splits (UAE vs KSA) rather than contradictory findings. Caveat: headline percentages can vary by how questions are framed (e.g., “have used AI” vs “use it all the time”), so precise policy or procurement decisions should reference the original questionnaire and sample frames in Marriott’s full methodology if granular accuracy is required.
Practical guidance for travellers who want to use AI safely
AI is a powerful travel assistant when combined with simple verification habits. Adopt a compact, repeatable checklist before acting on AI recommendations:
- Ask the AI for sources and timestamps for operational details (e.g., “What is the source for the ferry schedule and when was it last verified?”).
- Cross‑check departure/arrival times, park or ropeway hours and guided‑tour availability on the operator’s official page or by phone.
- Never paste full PII (passport numbers, credit card CVV, full booking references) into a public chatbot; use secure portals or enterprise tools with contractual non‑training guarantees.
Additional practical tips:
- Use paid tiers or enterprise offerings that explicitly state data handling and non‑training clauses when booking or handling company travel.
- For bookings initiated via AI, insist on written confirmations from the supplier (hotel, airline, tour operator) and retain those audit trails.
- Enable payment protections (cards with chargeback policies or travel insurance) and document provider contacts for rapid escalation.
What hotels, OTAs and travel tech vendors must do
- Implement RAG and live connectors to authoritative supplier APIs to reduce hallucination risks; surface provenance clearly in the UI.
- Offer non‑training and data residency contractual options for enterprise and high‑value customers, and make those guarantees explicit in product tiers.
- Preserve human checkpoints for safety‑critical or high‑value actions (final booking confirmations, refunds, complex itineraries) rather than fully automating closure without named human oversight.
For hospitality brands specifically, lean into lux‑scaping by packaging short, premium add‑ons that can be surfaced at booking time by AI assistants, but ensure inventory and pricing are synchronised in real time to avoid mismatch and reputational harm.
Regulatory and consumer‑protection considerations
Policymakers should urgently clarify consumer protections when transactions are initiated or completed by AI agents. Key regulatory priorities include:
- Audit trails and consent records for actions taken by an AI on a user’s behalf.
- Transparency requirements for provenance and verification status of operational facts embedded into AI outputs.
- Data‑handling standards that define when traveller data can and cannot be used for model training.
These measures will reduce ambiguity around liability, pressure on dispute resolution channels, and the potential for harmful outcomes driven by inaccurate recommendations.
WindowsForum readers: tips for integrating AI into travel workflows
Windows‑centric users and IT administrators should be mindful of the ecosystems in play:
- Microsoft Copilot integrates deeply with Outlook, Teams, and Windows reminders — useful for synchronising travel confirmations, calendars and corporate approvals. Use enterprise Copilot tiers where available to get stronger privacy and governance protections.
- If you rely on Google Workspace or Google‑centric workflows, Gemini offers export paths to Google Sheets and Google Travel; confirm how export data is stored and shared.
- For sensitive corporate travel arrangements, treat consumer chatbots as ideation tools only; use managed corporate travel platforms with contractual protections for bookings. IT procurement should require non‑training clauses and data residency guarantees before routing PII through third‑party models.
For Windows power users building personal planners, combine AI ideation with local syncing: draft itineraries in a Copilot session, export key items to Outlook/Calendar, and use Windows security best practices (BitLocker, up‑to‑date Defender definitions, and secure credential storage) when storing travel documents locally.
Longer‑term outlook: what to expect in travel tech
Expect rapid maturation along three axes:
- Action‑capable agents — scoped and permissioned AIs that can book, modify and refund when given explicit consent, backed by audit logs. Vendors are actively building these features.
- Deeper supply‑side integration — revenue engines and distribution systems will increasingly be built with ML‑native components (continuous pricing, personalized bundling), requiring new commercial and regulatory guardrails.
- Ubiquitous embedding — AI will be everywhere in the booking funnel, often unlabelled; the crucial question will shift from “does it use AI?” to “does it reliably work and can I trust the outcome?”
These developments can materially improve convenience and personalization — but they heighten the need for rigorous verification, privacy protections and clear lines of accountability.
Conclusion
Marriott Bonvoy’s 2026 Ticket to Travel findings capture a pivotal moment:
AI travel planning has moved from experiment to everyday tool for Gulf travellers, and trends like lux‑scaping and passion pursuits reveal sustained appetite for premium, interest‑driven experiences. The opportunity for travel brands, hoteliers and tech vendors is real: AI can increase inspiration, streamline planning and surface higher‑value purchases. Yet the confidence travellers express must be tempered with operational caution. Generative models still hallucinate, many consumer chat services retain data unless contractual protections exist, and the emergence of agentic booking systems raises fresh legal and consumer‑protection questions. Practical safeguards — provenance, human checkpoints, non‑training contractual options and explicit consent for agent actions — are not optional extras; they are foundational if travel AI is to scale safely and sustainably.
For travellers in the UAE and Saudi Arabia, the next 12 months will likely bring faster, more personalised trip discovery and an expanding menu of premium add‑ons. For industry players and regulators, the priority must be to translate that promise into reliable, auditable systems that protect consumers while letting innovation deliver on the clear demand Marriott’s research has identified.
Source: Arabian Business
Two-thirds of UAE and Saudi travellers trust AI to book holidays - report