AI Travel Agents vs. Direct Booking: Accor's H1 2025 Profit Warning

What happened

Accor reported growth in first-half profit despite a softening global travel market, demonstrating that disciplined revenue management and a diversified portfolio of brands can sustain margins even when broader demand cools. The French hospitality group's results arrived alongside a growing industry conversation about a structural shift in how travellers will book hotels in the near future.

That shift centres on AI-powered travel agents — autonomous booking tools embedded in platforms such as ChatGPT, Google's Gemini and a wave of start-up applications — which are moving from novelty to genuine commercial infrastructure. Rather than browsing and comparing options themselves, a growing cohort of travellers is delegating the entire search-and-book journey to these agents, which surface, filter and transact on the guest's behalf.

The emerging concern for hotel operators is that AI agents may become the dominant intermediary between a property and its future guests — a role previously held by online travel agencies (OTAs) such as Booking.com and Expedia. If that transition accelerates, hotels risk losing the direct-booking gains they have spent a decade fighting to reclaim.

Why it matters

For customer-experience and service-design practitioners, the rise of AI travel agents represents a profound change to the pre-arrival journey — arguably the phase where brand perception and emotional anticipation are first formed. When an algorithm selects and books a property on a guest's behalf, the hotel loses its earliest opportunity to shape expectation, communicate its distinct personality and begin the relationship on its own terms. The discovery moment, which has always been a latent CX touchpoint, is effectively outsourced.

From a behavioural-economics perspective, this matters because choice architecture is shifting hands. Hotels have long invested in website design, loyalty nudges and direct-booking incentives to influence the decision environment. An AI agent collapses that environment into a ranked output determined by the model's training data, pricing signals and optimisation criteria — criteria the hotel may not fully understand or influence. Operators who have built their CX strategy around owning the digital funnel will need to rethink what "owning the guest relationship" actually means when the guest never visits their website at all.

By the numbers

• First-half 2024: Accor reported profit growth despite a described softening in the broader travel market.

The Renascence take

Most operators will read this story as a distribution problem — a new channel to manage, a new algorithm to optimise for. That framing is understandable but dangerously narrow. The deeper issue is that AI agents are not just a new front door; they are a new curator of relevance, and relevance in that context is determined by structured data quality, review sentiment and price-to-value signals — not brand storytelling.

The hotels that will fare best in an AI-intermediated world are not those with the biggest marketing budgets, but those with the most coherent, data-rich identity — properties where every guest touchpoint generates signals that a model can read and reward. The contrarian move right now is to invest less in paid acquisition and more in the operational consistency that produces strong, specific, repeatable reviews. Loyalty programmes also need rethinking: if a guest never consciously chooses your hotel, the traditional points-and-perks model loses its behavioural hook entirely. Customer-obsessed operators should be asking today — what does our brand look like to a machine that has never felt hospitality?