ChatGPT IBE

Hotels and hotel brands lose bookings when the path from intent → availability → payment → confirmation is slow, fragmented, or requires users to navigate complex UIs.

• Conversion Friction: Traditional booking flows force users to hunt through filters, rate rules, and forms; drop-off happens before checkout.
• Support Load: Guests ask repetitive questions (“Do you have a double room tomorrow?”, “Is this refundable?”) that agents answer manually.
• Disjointed Systems: Inventory lives in the PMS, payments in a PSP, and the booking UI elsewhere—causing integration, UX, and reliability gaps.
• Brand Rollout Cost: Creating and maintaining multiple branded booking experiences is expensive (design, dev, QA, hosting).

A conversational, end-to-end booking experience inside ChatGPT.

This app connects ChatGPT to Apaleo (PMS/offers & reservations) and Adyen (payments) using an MCP (Model Context Protocol) server plus modern embedded widgets for a complete booking flow:

Offers → Cart → Payment → Booking Confirmation

Key differentiators:

• Real-time availability & pricing via Apaleo offers
• Secure payment checkout via Adyen Drop-in
• Dynamic guarantee handling (Prepayment / CreditCard / PM6Hold) based on each offer’s minGuaranteeType
• White-label build pipeline to generate per-brand assets/config and a deployable Docker image

1. Guest asks in natural language (e.g., “Show me available rooms in Munich for tomorrow, 1 adult”).
2. ChatGPT calls get_offers on the MCP server.
3. Apaleo returns offers (availability, price, guarantee type).
4. Offers widget renders a carousel of room cards (images + price + key conditions).
5. Guest selects a room; the cart/guest form is shown inline in the widget.
6. Guest submits guest details (name, email, address, phone).
7. ChatGPT calls create_payment_session; backend creates an Adyen session with correct metadata.
8. Guest completes payment in Adyen Drop-in.
9. Backend finalizes booking in Apaleo (with race-condition protection to prevent duplicates).
10. Confirmation widget displays booking details (and optional email confirmation can be sent when configured).

Inputs

• Search intent: property, dates, adults (via chat)
• Offer selection: chosen offer/rate plan from the offers widget
• Guest data: name, email, phone, address (via cart widget)
• Payment: card / wallet payment in Adyen Drop-in
• Brand config (white-label): colors, logo, supported properties, room images

Outputs

• Room offers UI: real-time priced offers, room images, conditions
• Cart UI: validated guest details entry + summary
• Payment session: Adyen Drop-in session + redirect / embedded checkout page
• Confirmed reservation: booking created in Apaleo PMS
• Booking confirmation UI: confirmation screen with key details
• Operational logs: structured JSON logs for debugging and auditability

• ChatGPT (MCP Connector): Chat-driven orchestration calling server tools.
• Apaleo: Offers retrieval + reservation creation (offers.read, reservations.manage, folios.manage).
• Adyen: Payment sessions + webhooks for authorization events.
• Email (optional): API-key based email sending (e.g. Resend) for confirmations.
• Deployment: Docker container + optional nginx reverse proxy and ACME companion for TLS.

Value for Guests (Conversion)

• Less friction: conversational search replaces complex filtering.
• Faster time-to-book: fewer clicks, guided flow, clear next step.
• Higher confidence: offer conditions and guarantee type handled correctly.

Value for Hotels / Brands (Revenue + Efficiency)

• Higher conversion rate through reduced UX complexity and guided checkout.
• Lower support cost by deflecting pre-booking questions into self-serve chat.
• Faster brand rollout via white-label builds instead of bespoke frontends.
• Reduced integration risk by standardizing the booking flow around Apaleo + Adyen.

KPI Targets (measure to prove impact)

• Conversion: search→checkout start, checkout start→payment success, payment→booking success
• Time-to-book: median minutes from first message to confirmation
• Drop-off points: offer view → selection, cart → payment, payment → confirmation
• Support deflection: fewer inbound “availability / price / refundability” queries
• Operational reliability: booking duplication rate (target: ~0), webhook error rate, API latency

• Guests increasingly expect conversational commerce and instant answers.
• Apaleo + Adyen provide robust APIs; MCP enables safe, tool-based LLM actions.
• White-labeling allows multi-property / multi-brand scaling without multiplying engineering effort.

• LLM Misinterpretation: Chat intent may be ambiguous; mitigate with tool schemas + validation and asking clarifying questions in-chat.
• Payment & Compliance:
  • PSP checkout stays within Adyen Drop-in, but you must still treat PII carefully.
  • Ensure GDPR/DSGVO handling (data retention, access, deletion) and legal review.
• Operational Dependencies: Availability and booking rely on Apaleo uptime and API limits; payments rely on Adyen availability.
• Webhook Reliability: Webhooks must be reachable and verified (HMAC) to avoid missed/forged events.
• Brand Configuration Risk: Incorrect mapping of room images/categories can reduce UX quality; requires content QA.
• Context & Session TTL: Session lifetime is finite; long conversations may need re-validation of availability and prices.