OTA & Travel Platform Strategy Intelligence Brief

Audience: OTA executives, metasearch leaders, travel marketplace strategists, platform product leaders Prepared from: The Phocuswright Conference 2025, San Diego Date: March 2026

Executive Summary

The Phocuswright Conference 2025 delivered an unmistakable signal: the OTA business model is not dying, but its operating assumptions are being dismantled in real time. The conference opened with Barry Diller — whose $1B+ Expedia bet in 2002 defined the OTA era — declaring that ChatGPT (900 million users) is "the first force capable of breaking Google's search monopoly in travel." That framing set the tone for two days of sessions in which Google's stranglehold on travel distribution was treated not as permanent, but as newly contestable. The question is not whether the top-of-funnel shifts — it is how fast, and who captures it.

The most concrete structural change discussed across multiple sessions is the emergence of AI as a parallel distribution channel. Google announced end-to-end agentic hotel booking within AI Mode during the conference week, naming Booking.com and Expedia as launch partners alongside hotel chains. Booking Holdings CSO Rob Ransom confirmed Booking.com had also launched a ChatGPT app via MCP two months prior. These are not pilot experiments — they are early bets on a channel that, per Phocuswright research, already influences nearly 40% of US travelers in trip planning. For OTA strategists, the operative question is whether AI channels prove additive or cannibalistic to existing traffic.

The financial analysts on the "Street Talk" panel offered the sharpest structural warning: even if AI doesn't disintermediate OTAs, if it routes more traffic back through Google — where OTAs pay approximately 80 cents per dollar of acquired traffic — the economics deteriorate sharply. Mark Mahaney (Evercore) simultaneously offered the most optimistic counter: ChatGPT is currently delivering high-quality leads to OTAs at zero cost. The short window in which AI platforms deliver traffic without charging for it is the most underappreciated opportunity discussed at the conference. OTAs that fail to establish strong AI channel presence before monetization models normalize will pay dearly later.

The platform competition is expanding from two primary directions: financial institutions building vertically integrated travel platforms (Chase Travel targeting $12B+ in gross bookings; iSeatz powering the $45-50B US financial institution travel market) and hotel chains using direct API pipes to AI discovery platforms to bypass OTA commissions entirely. Agoda CEO Omri Morgenshtern reframed the strategic imperative most crisply: OTAs should use AI to "move up the funnel and become the new travel search engines themselves," replacing dependence on Google. That is now both the threat and the opportunity at the center of OTA strategy.

Key Findings

1. Google's Monopoly Is Contested — and OTAs Have a Window

Barry Diller's opening session framed the competitive landscape in stark terms: Expedia and its nearest competitor alone spend over $10 billion per year on Google search. He argued that AI — specifically ChatGPT — is the first technology since Google's rise to meaningfully challenge that monopoly. The Wall Street panel amplified this: Mahaney noted Google search has visibly improved in the past two years due to competitive pressure from AI, and that Meta has now captured approximately 10% of performance marketing spend.

The Phocuswright consumer survey data adds urgency: nearly 4 in 10 US travelers now use ChatGPT, Google Gemini, or Perplexity to plan trips. At the top of the purchase funnel, traditional Google search is losing share to AI-powered tools, and review sites are declining simultaneously. This is a two-front erosion of OTA lead generation infrastructure.

The opportunity: Mahaney specifically noted that OTAs are currently receiving high-quality leads from ChatGPT at zero cost — a window that will close once AI platforms establish their paid distribution models. OTAs with early integrations (Booking.com's ChatGPT app, Expedia's Google AI Mode partnership) are acquiring free traffic that will eventually be monetized.

Implication: The next 18-24 months represent the lowest-cost window for OTAs to build AI channel presence before those channels impose distribution fees comparable to Google's.

2. Agentic Booking Is Live — But the Commercial Model Is Unresolved

Google's James Byers confirmed during his presentation that Google announced end-to-end agentic hotel booking within AI Mode during the conference week, with Booking.com, Expedia, Choice Hotels, IHG, Marriott, and Wyndham as launch partners. Byers described 2026 as a "year of learning" about what users actually want from agentic booking and predicted technical payment integration challenges would be resolved within months.

Rob Ransom (Booking Holdings CSO) discussed both integrations: the Booking.com ChatGPT app launched approximately two months prior via MCP-enabled queries, and the Google AI Mode partnership announced the Monday of the conference. He openly acknowledged that "the mechanics of how the two OTAs would be differentiated within AI Mode results — whether by consumer brand preference invocation or algorithmic decision — are still being worked out by Google."

Sabre CEO Kurt Ekert went further, launching Concierge IQ — an AI conversational commerce product deployed with Virgin Australia covering research, search, booking, and post-trip support across WhatsApp, airline.com, and agency channels. He positioned agentic AI as "a fundamentally new distribution channel, analogous to how OTAs emerged 30 years ago."

The critical unresolved question: How will OTAs be differentiated within AI-curated results? Brand recall invocation (consumers saying "book with Booking.com"), algorithmic selection, or bidding? The answer determines whether AI channels replicate Google's power dynamics or create new ones.

3. Hotel Chains Are Building Direct AI Pipes — Targeting OTA Commission Rates

The most strategically provocative session for OTA platform leaders was the panel with Richard Holden (former Google Travel VP), Stephen Kaufer (TripAdvisor founder), and Sanjay Vakil (CEO, DirectBooker). Vakil reported that DirectBooker — a low-cost API connecting hotel chains directly to AI discovery platforms like ChatGPT and Gemini — has signed contracts with six of the top ten hotel chains globally.

The value proposition is explicit displacement: DirectBooker is priced at a fraction of OTA commission rates, below credit card processing fees, while ensuring hotels retain merchant-of-record status, guest email and phone data, and loyalty point crediting. Holden noted that Google's own team found consumers "significantly more trusting and engaged when supplier-direct content was mixed alongside OTA content."

Meanwhile, Cloudbeds CEO Adam Harris reported LLM referral traffic to independent hotels is "growing exponentially," with the "super-majority going direct to properties." He predicted that within five years, "consumers will interact with brands directly rather than through websites, and the OTA intermediary layer will erode."

The countervailing data point: Kaufer argued this is not zero-sum — Booking.com has grown even as hotel chains increased direct booking share — and that AI may prove additive rather than displacing. Ransom echoed this, arguing the "narrative that Google will disintermediate OTAs is almost as old as Google itself."

4. Financial Institutions Are Building Competing Travel Platforms at Scale

The "Everyone's trying to sell travel" panel revealed a structural competitive threat that receives less attention than AI: financial institutions are deploying the one asset OTAs have failed to build — consumer trust — as a platform wedge.

Jason Wynn (CEO, Chase Travel) described Chase as building a "vertically integrated travel operating system" through acquisitions (CX Loyalty for booking/fulfillment, Frost for luxury/corporate, The Infatuation for dining, Fig for payments/offers) and airport lounges. Chase Travel is growing at approximately 20% CAGR and was targeting $12B+ in annual gross bookings. Phocuswright research cited the US financial institution travel market at $45-50B in total gross bookings — a significant and growing fraction of the overall market.

The loyalty dimension is acute: Phocuswright research found Gen Z and millennials report greater loyalty to non-travel brands than travel brands, with traditional loyalty tiers perceived as inaccessible. This cohort — 65% of Chase's new card acquisitions — is the precise demographic OTAs need to win.

Kenneth Purcell (iSeatz) noted a critical structural dynamic: while OTA B2C businesses face structural pressure from financial institution platforms, OTA B2B supply businesses (he cited Expedia's B2B segment, growing 25%+ YoY and representing ~30% of Expedia's $100B+ in gross bookings) benefit as the supply layer behind these platforms. This B2B bifurcation is central to how major OTAs should think about platform competition.

5. The Look-to-Book Explosion Under Agentic AI Creates Infrastructure Risk

Sabre CEO Kurt Ekert surfaced a technical distribution challenge that strategic planners need to understand: agentic AI will generate "logarithmically more search queries," dramatically worsening the look-to-book ratio that NDC has already pushed higher. He argued that intelligent caching at petabyte scale — a capability only a few players globally can provide — is becoming a critical competitive moat.

This has direct implications for OTA platform economics. Every AI agent conducting travel research on behalf of users will generate search load without a corresponding booking in most cases. The GDS and distribution infrastructure costs of serving AI-initiated queries at volume could erode margins for intermediaries who don't control caching infrastructure.

Separately, ARC's Jennifer Watkins noted agentic commerce was "absent from airline distribution conference agendas in spring 2025 but dominated every panel by late 2025" — signaling how rapidly the infrastructure readiness question has moved up the agenda. Airlines' primary concerns include disintermediation from the customer relationship, fraud exposure in new agent-initiated payment flows, and readiness of third-party distribution channels.

6. AI Search Disruption Is Reshaping Metasearch's Value Proposition — But Not Killing It

The Skyscanner/KAYAK/Civitatis panel challenged the most prevalent conference narrative directly. Bryan Batista (CEO, Skyscanner) pushed back on Agoda CEO Morgenshtern's claim that OTAs could become their own search engines, arguing "search is hard" and that Skyscanner's 20 years of experience and 160 million monthly travelers gives metasearch a durable position. He reframed AI as enabling "metasearch on steroids" — richer personalization, memory, and package aggregation — rather than obsoleting the category.

Paul Jacobs (GM, KAYAK) acknowledged that AI "does the opposite job of metasearch" — curation to narrow decisions — but argued customers ultimately still want price comparison and a trusted brand at the moment of booking. KAYAK's expansion into corporate travel via a Block Sky partnership is being positioned as the rational application of consumer-grade technology to a less-served segment.

The Wall Street panel added a sobering data point: Booking's write-down of Kayak was described as "a possible early casualty of generative AI's impact on travel planning," and TripAdvisor's organic traffic is in "meaningful decline partly attributed to generative AI." The distinction appears to be between metasearch platforms that have moved aggressively into AI-native experiences (Skyscanner, KAYAK) versus those that haven't.

7. The Personalization Gap Remains Unresolved — and AI Hasn't Fixed It Yet

Despite years of investment, the personalization gap persists. The "New Travel Seller Arena" panel produced a striking admission: all three panelists — Chase Travel, Hopper, and iSeatz — acknowledged that "despite access to vast customer data, meaningful personalization remains largely unrealized and is not materially better than it was in 2019."

American Airlines VP Neil Geurin explained why in concrete terms: AA has records on 40 million customers, but parsing that data meaningfully within the sub-second window a consumer will tolerate before abandoning a search is "technically unsolved." His near-term solution — focusing personalization on the post-booking, pre-travel window where milliseconds don't matter — is the pragmatic intermediate step most platforms should consider.

Agoda's Morgenshtern framed the product philosophy challenge as a hybrid that "nobody has cracked yet" — neither purely conversational chat nor purely improved UX. Agoda is running thousands of daily experiments to find the answer. Steve Singh (Spotnana) set the precision benchmark at 95% accuracy on the first recommendation — arguing most search engines deliver 95% irrelevant content and business travel AI must invert that ratio. He estimated the capability is two to three years away.

8. OTA B2B Is a Structural Growth Engine, Not a Fallback

Multiple sessions converged on OTA B2B as an underappreciated structural growth vector. The iSeatz/Chase panel noted Expedia's B2B segment is growing 25%+ year-on-year and represents approximately 30% of Expedia's $100B+ in gross bookings. Financial institution platforms, corporate travel tools, and the emerging agentic commerce layer all require wholesale supply — a segment where OTA inventory depth and API infrastructure create durable competitive advantages.

Hopper Technology Solutions' pivot is instructive: HTS now represents over 90% of Hopper's revenue by licensing fintech products (price freeze, cancel-for-any-reason, disruption assistance) via API to any travel seller globally. The Nubank Brazil case study illustrates the model: instead of loyalty points, Nubank offers travelers 12-month installment plans at 0% APR, a compelling proposition versus Brazil's 35% average consumer APR. HTS's agentic AI assistant is powering approximately 3 million live customer service conversations per year and is being deployed with a major Japanese bank.

Strategic Implications

The AI channel window is open and closing. OTAs currently receive traffic from AI platforms at zero cost. The cost structure will shift — either toward direct monetization models (as Google is beginning to implement through AI Mode partnerships) or toward algorithmic inclusion decisions that favor platforms with strong API integrations and content quality. OTAs should treat this moment as analogous to the early mobile web: the platforms that built native mobile experiences when mobile was still "experimental" compounded enormous advantages.

Distribution cost structure needs scenario planning. The near-term risk is not direct disintermediation but traffic migration back through Google under worse economics. If AI Mode pulls search queries back to Google's interface before routing to OTA booking, OTAs pay distribution costs on traffic they previously captured more cheaply through organic and social channels. Every OTA's planning assumptions should include scenarios for Google AI Mode becoming the dominant search interface at current or higher CPC rates.

B2C versus B2B strategic clarity is urgent. The financial institution platform threat and the corporate travel evolution both suggest that OTA B2C faces more sustained structural headwinds than B2B. OTAs with strong B2B infrastructure (API supply, fintech products, white-label capabilities) are better positioned to be the supply layer behind whoever wins consumer distribution — whether that is Chase Travel, a financial services platform, or an AI assistant. Expedia's B2B segment growth rate (25%+ YoY) suggests this thesis is already playing out in the market.

Supplier direct via AI is a real but slow-moving threat. DirectBooker's progress (six of the top ten hotel chains signed) should be monitored, but Kaufer's assessment that this is additive rather than zero-sum has historical precedent. The more acute risk is that OTAs lose the content depth competition: hotel chains have granular property data (hot tub temperatures, pet weight limits, member-only rates) that OTAs never captured, and AI platforms are hungry for exactly this type of data. OTAs need a content strategy for AI channels that goes beyond price-and-availability feeds.

Post-booking is the highest-ROI AI opportunity right now. Multiple sessions converged on post-booking operations as both the most immediate AI opportunity and the area of highest pain. Acai Travel's reported 60% reduction in call center handling time, Wenrix's 93% automation benchmark, and Amadeus's explicit call-out of travel disruption management as "the next major frontier" all point in the same direction. Post-booking AI has measurable ROI, shorter implementation timelines, and doesn't require solving the distribution channel question.

Action Items

Immediate (0-6 months)

1. Audit AI channel presence. Map current referral traffic from ChatGPT, Perplexity, Gemini, and AI Mode. Establish baseline before traffic becomes monetized. Prioritize MCP integration to enable richer API-level queries from AI agents.

1. Accelerate agentic AI product launches. Wall Street analysts (Mahaney) explicitly criticized Expedia's Romie and Priceline's Penny for "being stuck in beta too long." Consumer-facing agentic products need to be in full production, not extended pilots. Target conversational booking that covers research through post-trip support in a single flow.

1. Establish Google AI Mode commercial terms. Both Booking.com and Expedia are launch partners. Competitors must understand the inclusion criteria, commercial model, and differentiation logic within AI Mode results before it scales to full deployment. This is the most important near-term distribution negotiation.

Short-term (6-18 months)

1. Invest in content depth for AI channels. Property-level content that OTAs historically didn't capture (granular amenity details, contextual imagery, local context) is now a competitive necessity for AI channel relevance. Hotel chains are actively using this content depth to route bookings around OTAs via DirectBooker-style integrations. Supplier content programs should be redesigned around AI query matching.

1. Build or partner on post-booking AI automation. The ROI case for post-booking AI is proven and measurable. Acai Travel (minority Amadeus investment), Wenrix/DeepFlow, and similar platforms offer partnership or acquisition paths to 60-90%+ automation of complex servicing tasks. This reduces customer service costs while improving CSAT — a rare combination.

1. Develop B2B supply strategy for financial institution platforms. Chase Travel ($12B+ gross bookings), the $45-50B US financial institution travel market, and the growing number of non-travel brands adding travel verticals all require wholesale supply infrastructure. OTAs that position their B2B API as the default supply layer for this channel can capture distribution share without the customer acquisition cost of B2C competition.

Strategic (18-36 months)

1. Plan for infrastructure costs of agentic search volume. Ekert's warning about the look-to-book ratio explosion under agentic AI is a real platform economics risk. Model the cost impact of AI agents generating search queries at 10x or 100x current rates and ensure distribution technology investments address caching and query optimization.

1. Evaluate loyalty program redesign for Gen Z and millennial cohorts. Phocuswright research shows these cohorts have lower trust in travel brands and perceive traditional loyalty tiers as inaccessible. Financial institutions are winning loyalty with credit card points and 0% APR financing. OTAs need loyalty products that work for travelers who take fewer trips and have lower spend — not just the frequent flyer top tier.

1. Define the OTA's role in the agentic commerce payments stack. Visa's Trusted Agent Protocol (TAP) and Stripe's Agent Commerce Protocol (ACP) will create new payment infrastructure for AI-initiated transactions. OTAs that establish merchant-of-record status in agentic flows preserve customer relationship data, upsell opportunities, and pricing control. Platforms that cede merchant-of-record to AI intermediaries risk becoming invisible infrastructure.

Sessions to Watch

| Session | Speaker | Why It Matters | |---|---|---| | "Life? Travel? Count on it!" — Barry Diller returns after 20 years | Barry Diller (IAC/Expedia Chairman) | Framing session for the entire conference: OTAs spend $10B+/year on Google, and AI is the first credible challenger to that monopoly — sets the strategic context for every other session. | | Here's how Booking.com is playing to win in the next era of travel | Rob Ransom (Chief Strategy Officer, Booking Holdings) | The incumbent OTA's AI distribution strategy in plain language: ChatGPT app, Google AI Mode partnership, app-first conversion data, and the unresolved question of how OTAs get differentiated in AI-curated results. | | Travel search in the AI era according to Google | James Byers (Group Product Manager, Google Travel) | Google's own account of AI Mode's architecture, the agentic booking partnership announcement, and explicit advice to travel brands on content strategy for AI search — essential reading for anyone managing Google relationships. | | Reinventing hotel booking with AI | Richard Holden, Stephen Kaufer, Sanjay Vakil (moderated by Chris Hemmeter) | The supplier-direct AI channel thesis: DirectBooker has signed six of top ten hotel chains. Three veterans (Google Travel, TripAdvisor, product leadership) debate whether OTA disintermediation via AI is actually happening. | | Everyone's trying to sell travel — The New Travel Seller Arena | Jason Wynn (Chase Travel), Dakota Smith (Hopper/HTS), Kenneth Purcell (iSeatz) | Chase's $12B+ travel platform, Hopper's B2B pivot to 90% of revenue, and the $45-50B US financial institution travel market — the most important competitive threat to OTA B2C that is not AI. | | Street Talk: Finance, investment, M&A, IPOs in travel | Mark Mahaney (Evercore), Lloyd Walmsley (Mizuho), Emma Taylor (Barclays) | Wall Street's clearest-eyed view of AI's OTA impact: the free ChatGPT traffic window, Gemini-in-Chrome as a real-time comparison threat, the Kayak write-down as early AI casualty, and where investor capital is actually flowing. | | Competing in Travel's Next Era with Sabre's CEO Kurt Ekert | Kurt Ekert (CEO, Sabre) | The look-to-book explosion risk under agentic AI, Concierge IQ product launch, and the infrastructure economics argument for intelligent caching — critical context for OTAs managing distribution technology costs. | | How Agoda's CEO Omri Morgenshtern is pushing boundaries far beyond its home turf