Seventy percent of travellers trust hotel staff as a travel-planning resource; only 22% trust AI to build an itinerary tailored to their interests. Those two figures, from Hilton's 2027 Trends Report, frame what the company calls the "Travel Trust Equation" — and they set the commercial question for European operators: if AI now owns discovery, what exactly is left for the hotel to sell as its own differentiator?

Hilton's answer is trusted local expertise. The report distinguishes between using technology to surface options and relying on human judgement to make decisions. AI can compare hotels and summarise reviews efficiently; it cannot tell a guest which restaurant suits a quiet dinner or when a crowded attraction is worth visiting. What Hilton does not say is how much of that staff expertise currently exists on property, or what it costs to train and retain at a time of persistent labour shortages across European hospitality.

What does the research say?

Four separate studies point in the same direction, each with its own numbers.

• Mews surveyed 3,250 travellers for Hospitality Net. 89% wanted AI working behind the scenes or a balance of AI and human-led hospitality, not AI running the front desk. 75% said good hospitality depends on human staff. 59% wanted AI to absorb repetitive tasks so employees could focus on rapport and trust.
• Expedia Group surveyed more than 5,700 adults across the US, UK and India for its report "The AI Trust Gap: Why Travelers Will Continue to Choose Trusted Brands". 53% were comfortable letting AI suggest options. 66% would not trust AI to make a purchase or booking on their behalf.
• Peer-reviewed research published in Humanities and Social Sciences Communications, based on three experiments with 708 participants, found human recommendations more persuasive than generative AI for trips planned in the near future — partly because participants judged humans more credible.
• Skift and Wunderkind report 24% of travellers trust AI-generated recommendations, while 43% feel overwhelmed by AI's growing presence. Adoption, in short, has outrun confidence.

The pattern holds across sample sizes, markets and methodologies: travellers will let AI narrow the field. They will not let it close the deal.

What does this change for hotel operators?

The report argues the debate has moved past whether travellers use AI. The relevant question is which decisions they will delegate — and three roles emerge.

AI creates convenience: destination discovery, hotel comparison, review summarisation, initial itineraries. Human expertise creates confidence: interpreting individual preferences, reading local conditions, applying firsthand judgement. Human connection creates what the report calls memorable hospitality: empathy, accountability, and a staff member who takes responsibility when plans change.

For chains, the strategic implication is division of labour, not substitution. Expedia's 66% refusal to let AI complete a booking means the transaction moment still rewards human intervention — and that brands investing in conversational AI at the point of sale may be investing ahead of consumer consent. Mews' data supports the inverse bet: automate the back office, redeploy the hours saved to the front desk.

For independent and boutique properties, the finding on near-term trip planning cuts the other way. If travellers weighing imminent departures respond better to human recommendations, last-minute direct bookings are precisely where staff expertise converts — and where commission-free channels can be defended against AI-driven aggregators.

The gap in all four studies is execution. None measures whether hotels can actually deliver expert staff at scale, or what training investment converts a 70% trust rating into incremental RevPAR. Trust in staff is an asset on paper; monetising it requires payroll.

What happens next

Hilton will develop the 2027 Trends Report findings into operational positioning over the coming cycle, and competitors face the same sequencing decision: automate the repetitive 59% now, or wait while AI-native booking tools erode the discovery layer entirely.