Hotel & Hospitality Executive Intelligence Brief
The Phocuswright Conference 2025
Prepared for: Senior hotel executives, hospitality operators, and accommodation platform leaders Focus: Competitive dynamics, technology adoption, distribution strategy, and guest experience innovation
Executive Summary
The Phocuswright Conference 2025 delivered a clear signal to hotel and hospitality executives: the rules of distribution are being rewritten faster than at any point since the OTA revolution. Agentic AI — systems that take autonomous action, not just provide information — is emerging as a fundamentally new booking channel, and nearly 60% of travel businesses report already experimenting with or scaling it. Google announced end-to-end hotel booking directly within AI Mode in partnership with Booking.com, Expedia, IHG, Marriott, and Wyndham on the opening day of the conference. The question for hotels is no longer whether to engage, but whether they will show up as active participants or passive inventory in someone else's AI layer.
The distribution pressure on independent and mid-scale hotels is intensifying from two directions simultaneously. On one side, AI platforms like ChatGPT and Google's Gemini are becoming the new "front doors" for travel discovery — and hotels that fail to feed rich, granular content to these systems risk invisibility. Bonafide's research found that one in four LLM responses about a travel brand contains inaccurate or outdated information, with direct booking losses as the result. On the other side, new intermediaries are moving aggressively into the space: Airbnb is piloting hotels in New York and Madrid, financial institutions like Chase Travel are building vertically integrated lifestyle platforms targeting $12 billion in gross bookings, and agentic commerce protocols from Stripe and Visa are restructuring who holds the merchant-of-record relationship in AI-initiated transactions.
There is, however, a genuine and under-appreciated opportunity in this disruption. For the first time in two decades, the data advantage may be shifting back toward hotel suppliers. DirectBooker's Sanjay Vakil reported signing contracts with six of the top ten hotel chains globally to pipe rich supplier content — granular details like hot tub temperatures, pet weight limits, and member-only rates that OTAs never captured — directly into ChatGPT and Gemini. Adam Harris of Cloudbeds described LLM referral traffic to independent hotels as "growing exponentially," with the majority going direct to properties, bypassing OTAs entirely. The AI era may reward hotels that invest in deep content and direct connectivity more generously than the search era ever did.
Internally, the operational transformation is equally urgent. The Levee platform (winner of the Global Startup Pitch Scaleup competition) and Yanolja's cloud PMS both reflect a systemic reality: most hotels lack the foundational data infrastructure to power the AI systems they need. Robert Buckman of Amadeus was direct — hospitality IT is "the space most in need of a full tech reboot." Labor cost is the number-one line item on hotel balance sheets, supply of workers is not growing, and the front desk, housekeeping workflow, and post-booking service layers all remain largely manual. The conference consensus is that AI augmentation of frontline operations is not a future state — it is an immediate operational necessity.
Key Findings
1. Google's AI Mode Is Already Booking Hotel Rooms — With or Without You
Google announced live, end-to-end hotel booking directly within AI Mode, in partnership with Booking.com, Choice Hotels, Expedia, IHG, Marriott, and Wyndham. James Byers, Google's Group Product Manager for Travel, described AI Mode as having over 75 million daily active users and confirmed the intent to resolve remaining technical challenges on payments and experience integration within months, with 2026 framed as "a year of learning about what users actually want from agentic booking."
The strategic implication is immediate: hotels not connected to Google's feed infrastructure — with live pricing, availability, and rich content — are at risk of disappearing from the fastest-growing discovery channel in travel. Byers explicitly advised travel brands to maintain strong feed data as a foundation while surfacing "niche, human, and factual content" that can now match complex contextual natural-language queries, such as "heated pools" or "proximity to hiking trails," which cross-reference maps data, reviews, and video rather than relying solely on structured feeds.
Hotels that are named Google AI Mode booking partners — currently the major chains — gain a significant early-mover advantage in what is effectively a new OTA-equivalent channel, but at terms that are still being negotiated.
2. LLM Visibility Is Becoming a Primary Distribution Risk — and Most Hotels Are Unprepared
Layton Han, CEO of Bonafide.ai (named one of PhocusWire's Hot 25 Travel Startups for 2026), presented a study finding that approximately one in four LLM responses about a travel brand does not align with that brand's factual information. He illustrated the commercial consequence with a concrete example: Mandarin Oriental Hong Kong's premium limousine service was unknown to Google Gemini, costing the hotel potential high-margin bookings on every relevant guest query.
The problem is structural. LLMs are trained on public data and have no automatic mechanism to ingest current, granular hotel content — room-level amenities, package pricing, pet policies, accessibility details — at the specificity travelers now demand in natural-language queries. Bonafide reports its platform improves LLM alignment by up to 40%, which Han equates directly to a 40% improvement in booking opportunity.
Kristie Goshow (former CCO, Peregrine Hospitality) confirmed at the "How to compete for the modern hotel guest" panel that "most hotel teams do not yet have tools to measure their visibility on generative AI platforms" — meaning the bleeding is largely invisible to operators today.
3. Direct Booking via AI Is a Real and Growing Channel — But Requires Investment to Access
The "Reinventing hotel booking with AI" panel — featuring former VP/GM of Google Travel Richard Holden, TripAdvisor founder Stephen Kaufer, and DirectBooker CEO Sanjay Vakil — presented the most actionable distribution argument of the conference. Vakil reported that DirectBooker has signed contracts with six of the top ten hotel chains globally to create a low-cost API pipe delivering deep hotel content directly to ChatGPT, Gemini, and other AI discovery platforms. The model is priced at a fraction of OTA commission rates, below credit card processing fees, while preserving merchant-of-record status, guest email and phone data, and loyalty point crediting — meeting the definition of a true direct booking.
Adam Harris of Cloudbeds supported this data point independently, describing LLM referral traffic to properties as "growing exponentially," with the "super-majority going direct." He argued that within five years, consumers will interact with brands directly rather than through websites, and that MCP (Model Context Protocol) layers allow hotel brands to connect directly to AI agents like Gemini without a new intermediary.
Lennert de Jong of citizenM corroborated this from his own experience: he booked a local Austrian ski property directly through ChatGPT that "never would have surfaced on Booking.com" — precisely the kind of long-tail discovery benefit that favors independent operators willing to invest in content depth and AI connectivity.
4. The Corporate Event Opportunity Is Structurally Underserved — and Hotels Are Losing Efficiency
Healey Cypher, CEO of BoomPop (named Travel Innovator of the Year at Phocuswright Innovation Launch 2025), identified a structural inefficiency that directly costs hotels revenue: over 7,500 corporate off-sites and retreats are planned every day, hotels cannot be booked for more than 10 rooms online, and the resulting RFP process yields less than 1% conversion. Hotel sales teams spend 45 minutes responding to each RFP lead regardless of quality.
BoomPop's AI-native platform — built on MCP architecture, connected to 1.2 million vendors — processes approximately $100 million in annual event spend and has already demonstrated a 4.9/5 event rating. For hotels, the platform provides an AI-enriched RFP inbox that pre-drafts full responses (targeting a reduction from 45 minutes to 4-5 minutes per response) and gives global sales organizations portfolio-level visibility into win rates and competitive benchmarking.
The opportunity here is bilateral: hotels that integrate with platforms like BoomPop capture corporate group demand they are currently missing while reducing sales team overhead. The 450 brands currently on BoomPop represent early adopters of a channel that will scale.
5. Loyalty Is Misunderstood — and Hotels Are Optimizing for the Wrong Metrics
Phocuswright Research Manager Madeline List presented original research that should recalibrate how hotel executives think about their loyalty programs. The data was direct: for hotels, value and price dominate selection criteria, followed by cleanliness and amenities, with loyalty programs entering the top five — meaning programs are table-stakes enhancers, not primary drivers, for guests who already have a quality-baseline preference.
More provocatively, 84% of leisure travelers engaged in at least one loyalty "gaming" behavior in the prior year — opening credit cards for sign-up bonuses before canceling, flying to maintain status rather than out of preference, and booking on behalf of others to accumulate points. These behaviors represent program engagement without genuine brand loyalty.
The most insidious driver of brand defection, List found, is not dissatisfaction but novelty — the desire to simply try something new — appearing as a top switching driver across hotels, airlines, and OTAs. List's conclusion: "the winners will be brands that simultaneously maintain strong product quality and pricing, offer a sense of novelty, and deliver genuine value through their loyalty programs." Marriott's Homes & Villas data point supports this — 97% of bookers are existing Bonvoy members, often deploying large point balances on premium home rentals, recapturing guests who had been leaving the Marriott ecosystem for Airbnb-style stays.
6. Hotel Operations Technology Has a Foundational Data Gap That Blocks AI Adoption
Two sessions addressed the hotel operations technology problem from different angles, arriving at the same conclusion. Levee CEO Al Lagunas (winner of the Global Startup Pitch Scaleup competition) was blunt: "the foundational data infrastructure required to apply automation, robotics, or AI does not yet exist in hotel operations." His solution — a multimodal AI room inspection tool — is a wedge into a broader operating system for housekeeping, maintenance, inventory, and performance analytics. The pitch was grounded in his own family history: his mother worked as a hotel housekeeper.
Jean-Jacques Morin of Accor echoed this systemic diagnosis at the "How to compete for the modern hotel guest" panel, identifying poor data quality from siloed PMS, CRS, and other systems as the primary obstacle — not the AI tools themselves. Adam Harris of Cloudbeds described the current challenge as getting hotel data "into a form that can power agentic AI conversations" before any guest-facing personalization is possible.
Robert Buckman of Amadeus was the most direct: he described the hospitality IT sector as "the space most in need of a full tech reboot" and confirmed Amadeus is actively pursuing it, including a minority investment in Acai Travel for post-booking AI automation. NDC complexity has already pushed call center handling times up 20-30%, while experienced human agents are retiring faster than new ones can be trained — making back-office AI not a future luxury but a present operational necessity.
7. Airbnb's Hotel Entry and the Blurring Lines of Accommodation Categories Are Accelerating
The "Hotels, Homes and Beyond" panel delivered a competitive warning that hotel executives should take seriously. Airbnb CFO Dave Stephenson confirmed the rebuilt app (launched May 2025) is piloting hotels in New York, Madrid, and a third city. He also reported that nearly half of Airbnb experience bookings in Q3 were not connected to accommodation — meaning services like spa treatments, personal training, and fridge stocking via Instacart are functioning as standalone customer acquisition channels that build Airbnb relationships independent of any accommodation booking.
Steve Schwab of Casago noted that Airbnb lacks a traditional loyalty program, but Stephenson signaled one is "increasingly logical" as the platform rounds out its offerings. The Wall Street analyst panel confirmed Brian Chesky's public commitment to hotels has already triggered analyst upgrades, though debate exists about whether Airbnb has waited too long.
The Marriott-Sonder post-mortem was equally instructive: Jennifer Hsieh described how Marriott moved quickly to protect guests when Sonder faced liquidity collapse, underscoring that brand reputation and operational control matter in partnerships. The lesson for hotels considering platform partnerships: the financial risk of a partner's model is your guests' problem when things go wrong.
Strategic Implications
Distribution will bifurcate between AI-native hotels and the rest. Chains that have already signed with DirectBooker, connected to Google AI Mode, and invested in LLM content infrastructure will compound their advantages through 2026-27. Independent hotels and mid-market operators that delay will cede AI discovery to whoever curates their listing — OTAs, AI aggregators, or no one.
The merchant-of-record question is the most consequential distribution issue of the next 24 months. Agentic commerce protocols (Stripe's ACP and Visa's TAP) are being designed to preserve merchant-of-record status for the original seller — but only if hotels actively participate in the ecosystem. Hotels that distribute through OTAs or aggregators in AI channels will likely lose direct guest data, loyalty crediting rights, and the post-checkout relationship that enables upsell and cross-sell.
Loyalty programs need a product-quality foundation and a novelty layer to work. The Phocuswright research is clear: programs that simply accumulate points without delivering a meaningfully better stay are vulnerable to novelty-driven defection. Hotels need to invest in product quality first, then wrap loyalty around genuine experience differentiation — not treat programs as a substitute for either.
The labor and operations crisis is the floor, not a secondary concern. With hotel labor as the number-one balance sheet line item and workforce supply not growing, AI-augmented operations — from AI room inspection (Levee) to automated post-booking servicing (Acai, Wenrix) to AI-powered RFP response (BoomPop) — are not innovation experiments. They are cost control requirements. Hotels that treat operational AI as a 2027 initiative will face a structural cost disadvantage long before then.
Action Items
Immediate (0-90 days)
- Audit your LLM visibility. Run your property across ChatGPT, Google Gemini, and Perplexity for your ten most common guest query types. Identify the gap between what LLMs say about your property and what you actually offer. This audit requires no budget — only 90 minutes of staff time — and will surface your most urgent content priorities.
- Verify your Google feed data quality. If you are not already a Google AI Mode booking partner, confirm your pricing and availability feeds are live, accurate, and granular enough to support contextual queries (heated pool, pet-friendly, proximity to hiking). This is the baseline for AI discovery across Google's 75 million daily AI Mode users.
- Assess your merchant-of-record position in emerging AI channels. Map which of your current distribution channels preserves your merchant-of-record status, guest contact data, and loyalty attribution in an AI-initiated booking. Identify channels where that status is at risk.
Near-term (90 days – 6 months)
- Invest in LLM content depth, not just traditional SEO. Travelers now query LLMs with granular, specific questions your current website content does not answer. Build or commission a structured knowledge base covering every property detail — room-level amenities, pet policies, accessibility specifics, membership benefits, seasonal programs — and ensure it is accessible to LLM data ingestion endpoints. Consider platforms like Bonafide for systematic LLM alignment measurement.
- Pilot a direct AI booking channel. Evaluate DirectBooker or equivalent MCP-based API connectivity to ChatGPT and Gemini as a direct distribution channel. The economics — below OTA commission rates, below credit card processing fees, with full guest data and loyalty attribution — are compelling for any property currently paying 15-25% OTA commissions.
- Connect corporate event inventory to AI-native platforms. Evaluate BoomPop and comparable platforms for your group sales function. The 45-minute-per-RFP problem is a measurable cost center; AI-enriched RFP automation with portfolio-level win-rate tracking is available today.
Strategic (6-18 months)
- Begin or accelerate your hospitality IT modernization. The foundational data problem — siloed PMS, CRS, and operations systems that cannot interoperate — blocks every AI initiative downstream. Prioritize cloud-native PMS migration and data layer integration as the prerequisite investment for agentic guest experiences.
- Redesign loyalty around novelty and experience, not points mechanics. Audit your loyalty program against the Phocuswright research framework: is your product quality the primary driver of preference, or is your program compensating for a product gap? Identify where novelty-driven experiences — exclusive access, unexpected surprises, curated add-ons — can be injected into the loyalty journey to reduce defection to the "just trying something new" motive.
- Build a formal AI governance model for distribution decisions. Establish a cross-functional working group (revenue management, distribution, tech, legal) to evaluate every new AI channel, agentic commerce protocol, and platform partnership against a consistent framework: merchant-of-record status, guest data ownership, loyalty attribution, and brand representation quality.
Sessions to Watch
- “How to compete for the modern hotel guest” — Jean-Jacques Morin (Accor), Lennert de Jong (citizenM), Adam Harris (Cloudbeds), Kristie Goshow — The most directly relevant panel of the conference for hotel operators. Covers AI-powered guest communication, OTA disintermediation via LLMs, data fragmentation, MCP connectivity, and why citizenM sold its brand to Marriott for loyalty access.
- "Reinventing hotel booking with AI" — Holden, Kaufer, and Vakil with Chris Hemmeter — Former Google Travel VP, TripAdvisor founder, and DirectBooker CEO debate whether AI is fundamentally reshaping hotel distribution power dynamics. DirectBooker's signed contracts with six of the top ten global hotel chains make this a concrete rather than speculative discussion.
- “Hotels, Homes and Beyond — The Expanding Hospitality Arena” — Jennifer Hsieh (Marriott), Dave Stephenson (Airbnb), Steve Schwab (Casago) — Essential viewing for understanding how the competitive boundaries between hotels and alternative accommodations are shifting, including Airbnb's hotel pilot, Marriott's Homes & Villas loyalty data, and the Sonder collapse post-mortem.
- "Travel search in the AI era according to Google" — James Byers — The most important distribution session for chains and independents alike. Byers announces the Google AI Mode hotel booking partnership list (IHG, Marriott, Wyndham among named partners), explains the grounding data advantage, and provides direct advice on feed quality and content strategy for AI visibility.
- Levee — Global Startup Pitch Scaleup Winner (Al Lagunas) — The most operationally grounded startup presentation of the conference for hotel executives. Levee's multimodal AI room inspection platform addresses the foundational data gap in hotel operations and represents a category of "operations AI" that is early-stage but capital-efficient to deploy.
- Bonafide — PhocusWire Hot 25 Travel Startups for 2026 (Layton Han) — The clearest articulation of the LLM alignment problem and its direct revenue consequence for hotel brands. The Mandarin Oriental limousine service case study is worth examining for what it implies about your own property's AI visibility gaps.
- BoomPop — Travel Innovator of the Year (Healey Cypher) — For hotel group sales and revenue management teams: the most actionable demonstration of how AI is restructuring corporate event booking and the RFP process. The 45-minutes-to-4-minutes response time reduction is a measurable near-term ROI for sales teams.
- "Loyalty — Back to the Roots" — Phocuswright Research with Madeline List — Essential grounding for any executive responsible for a hotel loyalty program. The data on why travelers defect (novelty more than dissatisfaction), the role of product quality as a prerequisite, and the behavioral gap between program participation and genuine brand loyalty should inform every loyalty strategy discussion for the next 12 months.