The Problem
Users store photos across multiple drives, folders and cloud accounts. Duplicates pile up, and moving files feels risky. Existing tools often force a new structure instead of respecting the one users already have.
PhotoPeek – All your photos, one understandable library
The problem: photos are scattered across drives, folders and clouds, full of duplicates and inconsistent structures. PhotoPeek is my attempt to create a non-destructive, AI-assisted photo manager that gives power users control without breaking their existing setup.
UX Design Process
From scattered folders to a tested concept
For PhotoPeek I followed a structured UX process: from problem discovery and research, through ideation and prototyping, to user testing and heuristic evaluation.
Discovery & Research
I started with informal interviews, a short user survey and a content audit of real photo libraries. Affinity mapping and empathy maps helped me group pain points like "I don't know where my photos live" and "I am scared to move files" into a concrete problem statement.
Sketches & Prototypes
I used quick sketches and brainstorming sessions to explore different navigation models (by source, by time, by "virtual folders"). Low-fidelity wireframes and a Figma prototype allowed me to test tasks like "create a collection across two drives" before writing any code.
User Testing & Iteration
I ran moderated usability tests with a small group (casual users and a hobby photographer). Tasks focused on findability, duplicate handling and export workflows. The insights fed into revised flows and a heuristic review based on Nielsen's principles (visibility of system status, error prevention, user control and freedom).
From problem to product
Designing PhotoPeek around real workflows
This project combines UX methods (research, mapping, prototyping and testing) with a technical stack (C#, .NET, file system indexing and AI) to solve a very concrete problem: keeping large photo libraries usable over time.
Tools & Technologies
UX: Figma for wireframes and prototypes, Miro for empathy maps and user journeys, Google Forms for surveys, Notion for documenting insights. Build: C# / .NET for the desktop app, local file indexing for performance, and AI models for duplicate detection and similarity search.
Information Architecture
I mapped typical tasks (finding trips, client projects, family events) into a structure with sources, virtual collections and saved filters. This keeps navigation close to users' mental models instead of exposing technical details like file paths.
Usability Heuristics
I explicitly reviewed the flows against Nielsen's heuristics: visibility of system status (progress while scanning sources), match between system and real-world terms, user control and freedom (non-destructive operations), consistency in filters and labels, and error prevention when deleting or deduplicating photos.
AI-assisted duplicate handling
The AI duplicate finder is designed as a review step, not an automatic "clean up" button. Users see clear comparisons and can replace duplicates with a single virtual reference, reducing clutter without losing trust in the system.
Task-focused export & sharing
Export flows are built around real tasks from interviews: sending a zipped selection to a client, preparing images for a cloud share or moving a curated set to another drive. This keeps the UI focused on outcomes, not just file operations.
What I learned from user testing
I ran small, task-based usability sessions to validate the navigation, duplicate handling and export flows. Here are a few representative insights that influenced the design.
For the first time, I wasn’t worried something would happen to my originals. PhotoPeek made it clear that nothing gets moved, which finally gave me the confidence to import years of messy folders.
Hobby photographer
Visibility of system status & onboarding
Being able to build a collection from two different drives without moving anything feels safe. I always know I can undo actions.
Power user with multiple drives
Non-destructive model & user control
The duplicate suggestions were clear. I liked that the app asked me to confirm and showed both images side by side.
Casual user with messy library
Duplicate flow & error prevention
PhotoPeek as a UX case study
This project is still in development, but already demonstrates my end-to-end process: defining the problem, choosing the right tools, running discovery and user testing, and applying usability heuristics to improve the product.