AR Experience · Mobile Design

Architecture Discovery

A mobile AR concept that lets people explore architecture through their phone — discovering buildings, layers and stories in physical space.

Role
UX/UI Designer — Freelance Concept
Duration
4 Weeks
Platform
iOS — AR
Tools
Figma · FigJam · Zoom · Google Docs
KUBIK — AR architecture discovery app on iPhone

What is Kubik?

KUBIK gives users instant historical context by simply pointing their phone at a building. The app delivers micro-stories, AR overlays, then-vs-now visuals and audio guides optimized for exploring cities on foot.

The Problem

Visitors and locals pass historically significant buildings every day, but the stories behind them often remain invisible. Existing solutions, such as museum apps, plaques, or city websites, are:not optimized for mobile use on the move,Text-heavy and difficult to skim

The Goal

Design a fast, mobile-first experience that allows users to:instantly understand the historical context of nearby buildings, Explore AR-based reconstructions of the past,Save locations and create personal city tours

KUBIK Screen 1 KUBIK Screen 2 KUBIK Screen 3 KUBIK Screen 4 KUBIK Screen 5 KUBIK Screen 6 KUBIK Screen 7 KUBIK Screen 8
Research

As a freelance concept project, KUBIK started from a defined proto-persona and empathy map — assumption-based tools I used to focus the design on one clear scenario: someone standing in front of a building, wanting instant context without effort. These are hypotheses to be validated, not field findings — but they kept every design decision anchored in a real situation.

User Persona

A working student new to Cologne who loves discovering the city on foot, but finds existing history apps text-heavy and slow to use on the move.

KUBIK proto-persona — working student exploring Cologne

Empathy Map

Mapping what she says, thinks, feels and does surfaced the core need: instant, visual context without effort — no registration, no long reads, no standing around looking lost.

KUBIK empathy map — says, thinks, feels, does
Ideation

With the persona's needs defined, I mapped how users would move through the app — from spontaneously discovering a place to building and saving personal tours. The flow splits into two clear paths, keeping casual exploration and deeper engagement separate.

User Journey Flow

Two entry points shape the experience: Discover as visitor for instant, no-login exploration, and Favorites & Tours for saving places, building routes and revisiting Cologne's stories.

KUBIK user journey flow — discover as visitor and favorites & tours paths

Swipe to explore →

Design

With the flow in place, I moved from structure to screens — starting rough on paper, then building up to a full UI and a consistent style guide.

Wireframes

I began with low-fidelity paper sketches to test layout and hierarchy fast — the onboarding slider, place overview, the past-vs-present comparison and the audio-guide player all took shape here before any pixel work.

KUBIK low-fidelity paper wireframes

Low-fidelity paper sketches

Style Guide

The visual system is built for outdoor use: a warm orange accent that stays visible in sunlight, high-contrast dark text, and Poppins — chosen for its geometric structure and excellent readability on mobile screens on the move.

KUBIK style guide — Poppins typography and color palette

Typography & colors

Prototype

I built a clickable prototype in Figma covering the core flows — discovering a place, comparing past and present, and playing the audio guide. Try it yourself below.

Reflection

KUBIK gave me the space to own a full concept on my own — from defining who it's for to shaping every screen. Working from a defined persona and empathy map kept every decision anchored in a real situation: someone standing in front of a building, wanting context fast, without effort.

The real challenge was translating an AR experience into a flat design tool. It pushed me to find clear visual metaphors — the scan frame, the past-vs-present toggle — that communicate depth without real AR. I also learned to focus over feature count, splitting the app into two clear paths instead of piling on features.

The natural next step would be validating these assumptions with real people on Cologne's streets — testing whether the scan-to-story flow works without any explanation.