Theme
UX Flows & Information Architecture

UX Documentation · v1.0

User Journeys, Flow Diagrams & Information Architecture

t.Co. Personal AI Travel Concierge · David Castiel · HIT M.Design 2026

01

User Personas

Three representative personas defining the range of t.Co. users — from experience-seeker to adventure lover to business traveler.

✈️
Maya, 31
The Experience Seeker
Background

UX Designer, Tel Aviv. Travels 4–5 times a year, mostly solo. Values authentic local experiences over tourist traps. Tech-savvy but hates complexity in travel planning.

Goals
• Discover hidden gems beyond guidebooks
• Minimal planning effort, maximum experience
• Connect with local culture & food
Pain Points
• Overwhelmed by too many options
• Doesn't trust generic recommendations
• Hates rebooking when plans change
t.Co. Value
AGT-01 builds her taste profile in minutes. AGT-02 surfaces local gems tailored to her style. AGT-05 auto-adjusts her schedule on disruptions.
Solo Culture Food Mid Budget
🏔️
Daniel, 26
The Adventure Traveler
Background

Freelance photographer, always on the move. Prioritizes off-the-beaten-path destinations, outdoor activities, and budget efficiency. Plans last-minute, often changes plans mid-trip.

Goals
• Flexibility to change plans on the fly
• Find untouched, photogenic locations
• Stretch every dollar of his budget
Pain Points
• Rigid bookings that can't be changed
• Language barriers in remote areas
• Missing the best photo spots & timing
t.Co. Value
AGT-04 monitors his location for off-path suggestions. AGT-06 bridges language gaps in real time. AGT-05 re-routes instantly when weather changes.
Solo Adventure Photography Low Budget
💼
Noa, 42
The Business Traveler
Background

VP Product at a startup, travels 2–3 times a month for work. Has loyalty memberships everywhere. Occasionally extends business trips for personal leisure. Time is her scarcest resource.

Goals
• Zero-friction bookings & disruption handling
• Maximize loyalty points on every trip
• Seamlessly blend work & leisure time
Pain Points
• Flight disruptions ruin tight schedules
• Loyalty points expire or go unused
• No time to research leisure add-ons
t.Co. Value
AGT-09 optimizes her loyalty across programs. AGT-05 auto-rebooking on disruptions. AGT-11 proactively suggests the perfect bleisure add-on.
Business Bleisure Loyalty High Budget
Maya — Experience Seeker Daniel — Adventure Traveler Noa — Business Traveler
Trip frequency4–5× / year8–10× / year24–30× / year
Planning styleResearches a lotLast-minute, flexibleDelegated / automated
Budget$1,500–$3,000$500–$1,200$3,000–$8,000
Primary deviceMobile + webMobile onlyMobile + voice
Key agentsAGT-01, 02, 06AGT-04, 05, 06AGT-03, 05, 09, 11
Retention driverPersonalized discoveryFlexibility & languageLoyalty optimization

02

User Journey Map — Traveler Using t.Co.

The journey map follows a traveler from the moment they consider a trip through their return home, including emotions, touchpoints, and which agent is active at each stage.

DISCOVER PLAN BOOK TRAVEL POST-TRIP LONG-TERM ACTIONS Hears about t.Co. Opens app, starts conversation with Tico Answers questions about preferences, reviews itinerary Confirms bookings, receives travel pack, saves itinerary Asks for directions, translation help, schedule changes Views trip memories, rates experiences, shares highlights Receives next trip suggestions, loyalty optimization alerts EMOTION 🤔 Curious 😊 Engaged 😌 Relieved 🤩 Delighted 🥰 Nostalgic ✈️ Anticipating TOUCHPOINTS App / Exhibition Voice Interface Chat + Map Itinerary View Confirmation Email / Calendar Voice + Map Push Notifications Memory Journal Rating UI Suggestions Feed Loyalty Dashboard ACTIVE AGENTS AGT-01 Profiling AGT-01 + AGT-02 AGT-03 Booking AGT-04 Location AGT-05 Schedule AGT-06 Language AGT-07 Memory AGT-08 Feedback AGT-09 Loyalty AGT-11 Anticipation OPPORTUNITIES Reduce friction in onboarding Surface hidden gems, live map One-tap confirm, zero manual entry Proactive alerts before disruptions Auto-curated shareable journal Timing next trip for best deals Profile update loop — every trip makes t.Co. smarter
Before / During Trip Agents
Planning Agents
Post-Trip / Long-Term Agents
Continuous Feedback Loop

03

Conversation Flow — User ↔ Tico

How a typical conversation unfolds between the user and t.Co.'s core agent, from opening to itinerary confirmation.

USER TICO (AI) Hi! I'm Tico 👋 Where have you always wanted to go? greeting 1 "I want to go to Japan" 10 days · budget ~$3k destination + constraints 2 AGT-01 active "Love nature or cities?" "Solo or with someone?" preference questions 3 "Mix of both, solo" food lover · no crowds profile refinement 4 AGT-02 active Generating day-by-day itinerary on live map… live itinerary generation 5 "Can we add Kyoto on day 5?" refinement request user feedback 6 Updated! Ready to book? AGT-03 initiates reservation flow confirmation prompt 7 "Yes, let's go!" ✓ Trip confirmed ✓

04

Agent-to-Agent Communication Flow

How the Orchestrator activates and coordinates agents, what data flows between them, and the logical order of activation.

USER Voice / Text Input CORE Orchestrator LangGraph / Claude raw input AGT-01 Preference Profiling activate AGT-02 Itinerary Creation user profile JSON AGT-03 Booking & Mgmt confirmed itinerary AGT-04/05/06 During-Trip Agents live context AGT-07/08 Memory & Feedback AGT-09/10/11 Long-Term Agents Vector DB PostgreSQL profile / memory synthesized response
Primary data flow
Agent activation
Post-trip / async
Deferred / background

05

Information Architecture — System Data Structure

The complete data structure of t.Co.: user profile, itinerary, memories, feedback, and trip history.

USER id · name · locale PROFILE Preferences · Style Budget range · curr Interests tags[] · wt Constraints diet · mobility ITINERARY id · destination · dates Days[] date · theme Activities[] place · time · cost Bookings[] PNR · status MEMORIES trip_id · timestamp Photos[] url · caption Journal text · mood FEEDBACK ratings · signals Ratings place · score Signals time · skips TRIP LOG history · metadata Trips[] dest · dates Spend total · breakdown 🗄 Pinecone (vector embeddings) 🗄 PostgreSQL (structured data) 🗄 Redis (session cache) · S3 (media) continuous profile refinement DATA ENTITIES UserProfile { id: UUID travel_style: string budget_range: {min,max} interests: string[] constraints: object embedding: vector } Itinerary { id: UUID destination: string dates: DateRange days: Day[] total_budget: number status: enum } Memory { trip_id: UUID photos: Media[] journal: string highlights: string[] embedding: vector } FeedbackSignal { place_id: string rating: 1–5 time_spent: seconds skipped: boolean context: object }
Profile data (Pinecone + PostgreSQL)
Itinerary data (PostgreSQL)
Memory data (S3 + Pinecone)
Feedback signals → profile update