Collaborative Research

How hotel rooms find their guests — and who takes a cut. From SABRE's first mainframe in 1960 to Priceline's $133M acquisition that became a $170B company, to MCP servers making OTAs optional. Interactive timeline of 60+ events, system architecture diagrams for three eras, and M&A ownership analysis of the Booking Holdings vs Expedia duopoly.

• GDS
• OTA
• MCP
• M&A
• Timeline
• Architecture

A clause-by-clause analysis of the standard distribution agreements that Booking.com, Expedia, and major OTAs require properties to sign. Rate parity clauses that prevent hotels from offering lower direct rates. Last-room-availability requirements that force hotels to keep inventory open even when they want to close. Marketing opt-outs buried in appendices. Liquidated damages for early termination. This research identifies the most onerous terms, explains what each one actually costs the operator, and provides a negotiation playbook — including which clauses are negotiable, which are dealbreakers for the OTA, and what independent properties have successfully pushed back on.

• Rate Parity
• LRA
• Contracts
• Negotiation
• Legal

An empirical analysis of how disputes are resolved today: Booking.com's resolution center, Expedia's partner support, Airbnb's guest-favoring arbitration, and the credit card chargeback process. Data on win rates, resolution timelines, and financial outcomes for hotels vs. guests. Includes the Airbnb pendulum analysis — how the platform shifted from host-friendly to guest-friendly as supply became abundant, and the long-term consequences for host trust. Makes the case for why ADAPT-DRP's localized arbiter model produces structurally better outcomes.

• Disputes
• Airbnb
• Chargebacks
• DR-WG
• Arbitration

A technical and business analysis of how programmable settlement transforms hotel payments. Compares merchant model, agency model, pre-payment, and BNPL flows — and shows how an open protocol replaces all of them. Explores the micro-financing marketplace: local credit unions offering per-booking loans, community development funds underwriting travel, and the novel arbiter-as-financier model where the dispute resolver also funds the reservation, aligning incentives. Revenue projections for the new financial services layer.

• Payments
• Settlement
• Micro-Finance
• Arbiter-Financier

A UX audit of major OTA extranet portals documenting deceptive design patterns that increase commission leakage. Multi-step workflows that discourage rate adjustments. Bulk editing interfaces that are deliberately cumbersome. Monthly statements in formats that resist reconciliation. Early departure and partial refund corrections buried behind unnecessary clicks. Quantifies the annual cost of dark pattern-driven over-payment across the industry and proposes transparency standards for ADAPT-compatible operator interfaces.

• UX
• Dark Patterns
• Extranet
• Commission
• Transparency

How bedbanks like Hotelbeds, WebBeds, and RateHawk can operate as transparent intermediaries on programmable rails. Covers net rate contract encoding, rate leakage prevention through machine-enforceable resale terms, event-based block management (auto-release, DMO contributions), and why bedbank inventory needs to be discoverable by AI agents. Includes the case for local event wholesalers — the Memphis operator who pre-purchases 500 room-nights for CMA Fest across 10 properties with terms on-protocol.

• Bedbanks
• Wholesale
• Rate Leakage
• Event Blocks
• B2B

Every ADAPT protocol — discovery, settlement, dispute resolution, guest bonding — converges at the property management system. This research examines what it means for a PMS to be protocol-native: guest history that travels with the guest, building access control tied to reservation state, housekeeping workflows triggered by programmable settlement events, and a single interface that publishes to AI agents, bedbank wholesalers, localized arbiters, and the entire visitor economy marketplace. Includes a technical case study of ExchangePMS — the reference implementation running on 200+ units at The Exchange Building.

• PMS
• Building OS
• Guest Journey
• Access Control
• Operations

Google Chrome 146 introduced native MCP support, enabling AI agents to interact with web content through the browser itself — reading pages, filling forms, executing transactions, and ingesting structured data without scraping or brittle DOM selectors. This is not an isolated feature. It is the latest signal in a pattern that is reshaping every layer of the technology stack: browsers become agent-aware, operating systems expose tool-use APIs, payment networks add agent settlement protocols (AP2, UCP), and property management systems publish MCP servers. The entire stack — from silicon to browser tab — is converging on a single assumption: AI agents are first-class users. This research traces the pattern across layers (hardware, OS, browser, network, application, commerce) and argues that accommodation providers who treat agentic compatibility as optional are building on an architecture that is already being replaced beneath them. The window to ADAPT is measured in browser release cycles, not decades.

• Chrome 146
• MCP
• Agentic Browser
• AI-Native Stack
• Infrastructure

Between January 29 and March 20, 2026, the hotel industry crossed from conference-deck agentic AI to live infrastructure. Accor launched the first major hotel ChatGPT app. Hyatt, Marriott, and Hilton each chose structurally different strategies — open platform, negotiated lane, and branded concierge. Lighthouse shipped the first MCP-based ChatGPT app open to any hotel at a flat fee. Aven embedded MCP across 35,000 SynXis properties. And OpenAI abandoned in-chat checkout entirely, ceding transactions to partner apps while Google pushed ahead with tiered agentic booking via Universal Commerce Protocol. This article reconstructs the six-week sequence, analyzes what each move reveals about platform economics, and identifies the three parallel booking models now emerging: AI-enhanced search, AI-assisted booking, and autonomous agent commerce. Includes the Skift/McKinsey data showing 124% YoY growth in extensive AI trip planning, the Morgan Stanley thesis that OTAs may benefit short-term, and why the real battleground is discovery — not transactions.

• ChatGPT
• MCP
• OpenAI
• Google
• OTA
• Discovery
• March 2026

In the GDS era, hotels connected to distribution via EDIFACT messages over private networks. In the OTA era, they connected via XML APIs and channel managers. In March 2026, a new pattern emerged: hotels publishing structured data via Model Context Protocol servers that AI platforms consume directly — no scraping, no inventory duplication, no intermediary markup. This research provides a technical and strategic analysis of MCP in hotel distribution. Covers the protocol architecture (tools, resources, prompts), how Aven's SynXis integration works for 35,000+ properties, how Agentic Hospitality's TravelOS MCP Server connects existing CRS/PMS to AI platforms without replacing infrastructure, and how Lighthouse built the first zero-commission ChatGPT app on MCP. Compares MCP against Google's A2A Protocol (agent-to-agent negotiation), Universal Commerce Protocol (AI purchases), and ACP. Includes a protocol decision matrix for hotel operators and PMS vendors, the ADAPT position on why MCP endpoints should be an open standard rather than proprietary vendor lock-in, and a technical roadmap for protocol-native PMS integration.

• MCP
• A2A
• UCP
• Protocol
• CRS
• PMS
• Aven
• Lighthouse
• Technical

The infrastructure is being built. The protocols exist. The platforms are live. But the consumer isn't ready. Skift's March 2026 data shows 30% of travelers now use AI 'extensively' for trip planning — a 124% increase year over year. Yet autonomous booking remains a trust fall most travelers won't make. This research examines the structural gap between technology readiness and consumer adoption. Analyzes three friction points: delegation anxiety (letting an agent spend money), preference opacity (AI can't yet capture the nuance of 'I want a hotel that feels like a boutique but isn't'), and accountability ambiguity (who is responsible when an AI agent books the wrong room?). Includes Hospitality Net's framework of three parallel booking models — and argues that the autonomous agent-to-agent future (A2A, delegated payments, zero-UI commerce) requires solving the trust problem before the technology problem. Proposes that ADAPT's localized arbiter model, guest trust deposits, and transparent intermediation may be the structural prerequisites for consumer trust in agentic booking — not just technological improvements in AI reasoning.

• Consumer Trust
• Adoption
• A2A
• Behavioral
• Arbiter
• Guest Trust Deposit

The strategic question isn't which protocol wins. It's who controls the structured data layer between your property and the AI agent. If it's a vendor, you've changed landlords. If it's an open protocol your PMS implements directly, you're free. This research compares three distribution interfaces — CLI (developer agents, automation), REST/GraphQL APIs (OTAs, metasearch, function calling), and MCP (AI-native discovery) — not as competing standards but as parallel adapters over a single canonical data layer owned by the hotel. The key insight: hotels that structure their property data correctly — Schema.org markup, clean rates, machine-readable policies, multilingual descriptions — can speak every protocol simultaneously. CLI for developer agents. API for traditional channels. MCP for ChatGPT, Claude, and Gemini. x402 for programmable settlement. Future protocols TBD. The PMS adds an adapter; the hotel doesn't change vendors. Includes the protocol comparison matrix (setup, standardization, AI reach, auth maturity, governance), the emerging Google stack (A2A + UCP + AP2 + Chrome 146 MCP), the x402 payment rail that eliminates 2.9% card fees, revenue retention math at Exchange Building scale ($272K/year recaptured), a competitive gap analysis of proprietary MCP vendors (Agentic Hospitality, 1stay.ai, Cendyn) vs. ADAPT's open protocol approach, and the five-step independent hotelier playbook.

• CLI
• API
• MCP
• x402
• Data Ownership
• Protocol-Agnostic
• PMS
• Revenue Retention