Crop.photo — AI Redesign · Process
The research, personas and flows behind Crop.photo — AI Redesign. The case study covers the decisions; this is the work underneath them.
02 — User Research
Three distinct user types driving demand for AI-powered image editing — each with unique workflows, constraints, and expectations from a tool like Crop.photo.
03 — Business Challenges
Users spent hours on repetitive cropping and resizing tasks that AI could execute in seconds — making the status quo actively harmful to business output.
Every major platform demands different image dimensions. Without batch resizing, each post required manual re-exporting — a compounding time tax on every content creator.
Complex product backgrounds — shadows, reflections, multi-tone surfaces — demanded expert-level masking that non-designers couldn't reliably execute at scale.
Fast tools sacrificed quality; quality tools sacrificed speed. Users were forced to choose between professional output and practical throughput — AI was the only way to deliver both.
04 — Secondary Research
78% of e-commerce sellers still crop and retouch product images manually — a massive unaddressed opportunity for AI automation at scale.
The average content creator wastes 45 minutes per day on image resizing and formatting tasks — nearly 4 hours of recoverable productive time per week.
AI-powered image editing tools reduce total editing time by up to 85% compared to manual workflows — validating the core product thesis and market timing.
07 — User Flow
08 — Toolkits
Tools and methods used throughout the design process — from initial research and wireframing through to final prototype and handoff.
A five-phase process covering both the marketing homepage and the application interface — grounded in e-commerce user research, competitive analysis, and conversion-optimised visual storytelling.
Audited crop.photo against Remove.bg, Canva, and Clipping Magic — studying homepage value communication, feature discoverability, and batch processing UX patterns. Mapped crop.photo's 8 AI features against how they surfaced in the existing interface.
Deliverables: Competitive matrix, feature audit, UX gap analysis.
Mapped three primary user archetypes — e-commerce sellers (bulk crop & background removal), marketing designers (multi-platform export), and fashion brands (headless model shots). Identified the marketplace export friction as the highest-priority UX failure point.
Deliverables: Persona cards, pain point prioritisation matrix.
Designed a before/after AI slider as the homepage hero — replacing the existing feature list with a live demonstration. A single drag communicates AI bulk cropping capability without any reading, targeting a 3-second value comprehension window.
Deliverables: Hero interaction model, demo component, CTA hierarchy.
Redesigned the core application screens: unified batch dashboard, marketplace-aware preset selector, AI confidence score per image, and dedicated entry points for Headless Face Cropper, Listing Analyzer, and AI Fashion Model Generator.
Deliverables: App wireframes, feature flow maps, interaction specs.
Produced pixel-perfect screens in Figma — homepage, batch dashboard, preset selector, and batch preview grid — with complete design system tokens, brand-accurate blue palette, and a component library covering all 8 AI feature surfaces. Crop.photo's users include professional e-commerce teams operating under accessibility requirements for their own storefronts. All components were built with WCAG AA as a baseline. Keyboard navigation paths were mapped for the batch upload flow — a particularly complex interaction — ensuring the tool is usable without pointer interaction.
Deliverables: Hi-fi prototype, component library, style guide, accessibility checklist, Behance publish.