CONTEXT-AWARE TRAVEL PLANNING
Designing a conversational map that helps people make contextual travel decisions
University project
MapAI
ROLE
Concept, Research * UI Lead
EXPERTISE
UX / HCI / AI Interaction
Date
February 2026
Usually when we plan a trip, it often means switching between maps, reviews, recommendations, and scattered sources of information. MapAI explored how a conversational interface could bring these decisions into one contextual map experience.
The system combines user preferences, travel constraints, and contextual information to generate personalized recommendations while keeping the user in control of the final decision.
Timeline
February 2026 · Academic project supervised by Janin Koch
Team
4 designers · Concept & UI/interaction led by me
We explored whether an AI assistant could help users express preferences naturally, reason about constraints, and visualize recommendations directly within the map.
The Problem
Fragmented information
Travel decisions require information from multiple sources.
Context matters
Recommendations depend on time, transportation, location, preferences and trip context.
Decision overload
Users need to compare options rather than simply receive another list of places.
Research & Planning
We conducted 8 user story interviews and explored how people currently make travel decisions.
Concept Development
We explored multiple interaction models including moodboards, map-based input, category selection, contextual prompting and AI-assisted itinerary building.
We started with paper prototyping.
I translated the concept into a high-fidelity Figma prototype and used a Wizard-of-Oz approach to simulate the AI responses.
Testing & Optimization
Three participants tested the prototype and identified problems around: AI discoverability, competing filters, user trust, location comparison, and contextual friction.
We then redesigned and re-iterate the flows of the user experience accordingly.
The AI shouldn't repeat the map. It should explain what the map can't.
Context-aware search
Users establish an origin, transportation mode and time/distance constraints before searching.
Conversational preferences
Instead of forcing users to translate their intent into filters, MapAI can ask questions and help clarify what they want.
Personalized recommendations
The system generates relevant places based on the user's preferences and contextual constraints.
Explainable recommendations
Recommendations communicate why something matches the user's needs.
Your testing found that participants appreciated AI explanations such as why a restaurant was recommended.
User-controlled AI
The user remains the decision maker. AI proposes, explains and adapts, rather than automatically deciding.
Basically, we let user be the "pilot" who sets constraints and makes qualitative decisions, while the system processes complex data and proposes options.
Users found AI explanations useful
Participants appreciated understanding why a restaurant was recommended, particularly when the explanation surfaced information such as dietary compatibility.
Users found the map intuitive
Participants understood the visual information on the map and naturally explored the recommended pins before interacting with the AI assistant.
Users struggled to compare options
Participants found it difficult to compare saved restaurants side-by-side, leading us to introduce a dedicated comparison view.
Users expected more contextual AI support
When information such as ratings and prices was already visible, users wanted the AI to provide deeper, personalized context rather than repeat existing information.
One month after we completed MapAI, Google introduced Ask Maps.
When we completed MapAI in February 2026, we had no knowledge of Google's upcoming product direction. One month later, Google announced Ask Maps, a conversational Maps experience that combines natural-language questions, personalized recommendations and map-based visualization.
The interaction model showed a striking convergence with the direction we had independently explored: users describe what they need conversationally, the system interprets contextual preferences, and recommendations are surfaced directly within the map.
MapAI
Conversational requests
Contextual preferences
Recommendations on map
Iterative exploration
Google Ask Maps
Complex natural-language questions
Personalized recommendations
Customized map showing options
Follow-up questions and actions