Is this the golden age of travel product or just another hype cycle? Well, the panel, comprising Booking.com’s David Adamczyk, Google’s James Byers, and Klook’s David Liu did not hesitate – it’s a golden age, they declared at WiT Singapore.
AI has raised the bar and moved the starting line. For the first time, product teams can answer real traveller questions in natural language, stay with users across the journey, and mine far richer behavioural signals at scale.
But that same power introduces new risks – of trust, of whiplash-level change, and of product fragmentation.
Below are the big challenges and the biggest opportunities called out by three builders in the arena.
Challenge #1: The slop problem (and the trust moat)
Risk: An internet flooded with low-quality, AI-generated content erodes both user trust and the data that trains future models.
Opportunity: Brands can become badges of quality – curating sources, tracking content lineage, and investing (heavily) in “ground truth” data: maps, schedules, hotel metadata, imagery, and verified human experiences. Trust compiles slowly and decays instantly; product leaders should treat it like a balance sheet item, not a campaign.
What teams are doing
- Google: Investing in curated, maintained datasets and long-game accuracy.
- Booking.com: Positioning brand as a seal of approval; moderating content and surfacing human-meaningful labels (e.g., “near Eiffel Tower” vs “7th arrondissement”) sourced via AI.
- Klook: Leaning on creator communities for authentic, on-the-ground content, then amplifying it.
Challenge #2: Pace without breaking the covenant
Risk: AI’s speed tempts reckless shipping; privacy and expectations can’t be collateral damage.
Opportunity: Bedrock principles (user trust, privacy, ecosystem value) paired with faster learning loops. Expect Google and others to test more imperfect ideas, kill them quicker, and scale the winners without weakening fundamentals.
Challenge #3: Vertical realities (and regulation)
All travel isn’t created equal.
- Accommodation: Enormous diversity (Accor to “John’s apartment”) plus a thicket of global regulations (consumer protection in California, payments in Indonesia, platform rules in Europe). Booking’s approach: platformise compliance so product squads can focus on creativity.
- Flights: AI can now reason over constraints, dates, and “vibes.” Google’s new Flight AI Deals lets you type “a week in a European foodie city in March/April” – and it finds real, discounted options. Science fiction, yesterday.
- Tours & Activities: Discovery remains brutally hard. Supply is endless, intent is fuzzy, and trips are multi-stop, multi-day. Klook is doubling down on niche supply and better intent understanding to match “first-time Japan” vs “done-it-all Japan.”
David Liu, Chief Customer Experience Officer, Klook
Challenge #4: Service at the point of disruption
When something breaks, seconds matter. Today’s “connected trip” still feels disconnected in customer service, especially for intermediaries stitching data from airlines, hotels, and operators.
Near-term plays
- Self-service with context: “It’s raining in Paris; here’s a family-friendly place on your route, no wait.”
- Agents assist with empathy: Klook built an emotion detector that flags sentiment to human agents in real time so they can de-escalate or delight, not just deflect.
Challenge #5: The “pinball problem”
Travelers still juggle tabs, notes, and calendars to plan. Builders see solutions within reach, but users say planning remains their top pain point. The funnel isn’t linear; it’s a pinball table. Winning teams must design for constant back-and-forth between inspiration, research, and booking across apps and moments.
Product mindset shifts
- Abstract constraints; localise creativity. Centralise model selection, governance, and compliance; decentralise AI experimentation to domain squads.
- Optimise for speed of learning, not volume of launches. More bets, smaller bets, faster feedback.
- Meet the traveller (and partner) where they are. Problems, not org charts, dictate roadmaps.
- Design for collaboration, not control. The connected trip will be solved by interoperability – data standards, shared context, and partnerable APIs – not a single walled garden.
James Byers, Group Product Manager Search, Google
What will likely get solved by 2030
- Itineraries: From laborious assembly to dynamic, context-aware trip scaffolds.
- Menus and micro-translation: LLMs crush this class of problem; friction vanishes.
- Operational logistics: QR codes, tickets, confirmations, and routing fade into the background.
What likely persists
- Flight delays. Physics and operations don’t optimize as fast as software.
- FOMO. Even perfect planning can’t cure human psychology – products must help users feel confident enough.
And in 2045?
We may still be talking about AI but the headline themes will be authenticity, privacy (always-on devices), and the meaning of being human in a world of agentic software.
David Adamczyk, Vice President, Product, Booking.com
A Playbook for product leaders (now)
- Build a Trust Stack
- Provenance: track sources and transformations of content.
- Verification: pair AI with human review where it matters (safety, legality, scams).
- Transparency: expose why a result is shown (signals, creators, recency).
- Instrument for Intent
- Move beyond keywords to capture missions (“first-time Tokyo with grandparents”) and constraints (budget, mobility, weather, mood).
- Treat intent as a first-class object that feeds discovery → conversion → service.
- Design for Nonlinearity
- Kill the one-way funnel. Offer save-and-resume, cross-device memory, and multi-party planning.
- Replace “filters” with natural-language refining (“somewhere warm, art + coffee, 4-hour flight max”).
- Make Service a Product Surface
- Build “interrupt-ready” flows: one-tap rebooks, proactive alternatives, contextual vouchers.
- Give human agents AI copilot UIs (sentiment, suggested responses, policy lookups).
- Platformise Compliance
- Central services for payments, identity, and regulation so feature teams can ship safely, fast.
- Win with Real-World Supply
- Treat creators and operators as partners, not inventory. Co-create, verify, and reward distinctive experiences.
- Measure Confidence, Not Just Clicks
- New North Stars: time to confident choice, plan completeness, post-trip satisfaction, service recovery NPS.
- Prepare for Data Collaboration
- Pilot shared schemas and privacy-safe data unions with key partners. The connected trip won’t happen without them.
Bottom line
AI has opened a wide new playing field, but trust and interoperability will decide who wins. The teams that pair rigorous curation with rapid learning and design for a pinball world will turn today’s hype into tomorrow’s habit.
