Year:
2026
Copilot
X Mariott
Project overview
Show how Marriott’s Bonvoy loyalty ecosystem can natively embed into Microsoft Copilot’s conversational interface, driving high-margin direct bookings, lowering distribution costs, and establishing Marriott as the premier "grounded" recommendation in AI-led decision journeys.
-
My Role: AI-Native UX Designer & Prototyper
-
AI- Native Workflow (48 Hours)
-
Tools: VS Code, GitHub Copilot, Copilot UI Skill
*This project is currently in an experimental pilot phase with limited access for initial testing only.
We targeted high-value, frequent business travelers who prioritize speed, convenience, and loyalty recognition.
-
The Core Problem: Business travelers want to quickly rebook past hotels without searching through old emails or going through a multi-step filter process.
-
The Target Persona: Alex (38) — Frequent Solo Business Traveler & Bonvoy Titanium Elite Member
-
Motivations: Wants maximum efficiency; wants to instantly utilize his massive bank of Bonvoy points and see his elite status perks.
-
Conversational Trigger: "Hotels in Seattle"
%20(1).png)
Core User Journey
When Alex inputs his request, Copilot uses secure Connectors to act as a reasoning bridge, delivering a frictionless booking experience:
1.Query Parsing & Real-Time Grounding:
Copilot parses the query to identify intent (Recently stayed), destination (Seattle), and payment preference (Points). It uses Marriott's structured, machine-readable data (amenities, reviews, real-time pricing) to ground the response with accurate, citable information at inference time.
2. Identity, Auth, & Connector Flow:
-
If Alex is not signed in, a secure Marriott Bonvoy sign-in flow pops up natively within the L2 canvas, avoiding external browser redirects.
-
Once authenticated, the Copilot Connector immediately displays his elite Titanium benefits (e.g., complimentary lounge access, available room upgrades) alongside live points pricing.
3. Checkout
-
Alex completes his booking securely right inside the L2 task page, preserving complete brand continuity to the final click.
Business Outcomes
We aim to kick off a strategic partnership exploration with Marriott to make Marriott the best-in-class hotel brand inside Copilot-led travel experiences, by enabling discovery, trust, personalization, loyalty, and checkout directly within Copilot-powered conversations.
The working web-based demo allowed stakeholders to witness a seamless direct booking flow in real-time, instantly shifting executive feedback from "Is this possible?" to "How soon can we ship this?".
Reflection: The 48-Hour "Vibe Coding" Workflow
To make this pitch irresistible to Marriott executives, a static deck wasn't going to cut it. I bypassed Figma entirely—trading vector-drawing for active building to craft a working, responsive prototype in record time.
-
Used conversational chatbots to rapidly talk through the interaction logic, generate state-management scripts, and plan the L1-to-L2 transition states.
-
Native Implementation: Leveraged the official Microsoft Copilot UI System directly in VS Code to build the interface, ensuring 100% compliance with native canvas guidelines.
before
