Product design · AI

RoomMind AI: redesigning a room from one photo in an iOS app

Photograph a room, pick a style, get an AI redesign in seconds. The full iOS app, from user flow to final UI.

RoomMind AI interior design app case study cover: the iOS home screen with New Design and recent projects
Project
Designed by Maniruzzaman Jubayer
Industry
Home and interior, AI, AI products work
Topics
AI product case studies, Mobile app case studies
Year
2026
Tools
Photoshop, Illustrator, XD, Stock, Figma
Featured on
Product design, Austin, Product design, Dhaka

Overview

RoomMind AI is an iOS app that redesigns a room from one photo, and this AI interior design app case study shows how: pick a style and get an AI redesign in under 30 seconds. I designed the app end to end, from user flow to final UI.

The deck names Europe as the priority market: the UK, Germany, France and Sweden.

The challenge

Most people can't picture a new look in their own room, and the app had to close that gap.

  • People can't see how furniture, colour and decor will look in their space, so they hesitate or make costly mistakes.
  • Inspiration is spread across several platforms, and planning by hand is slow.
  • The wait while the AI works had to feel guided, not blank.

The persona, Amir Rahman, an aspiring home designer, wants professional ideas without a designer's cost or time, and confidence before buying anything.

The process

  1. Research and market analysis, a problem and solution map, and a persona.
  2. A user flow from onboarding to the generated result, with branches for the AI scan, advanced options, errors and the profile.
  3. Low fidelity wireframes for every key screen.
  4. A style guide: Geist in four weights, a red accent, black, two greys and white.
  5. High fidelity UI, AI testing and validation, user testing and handoff, over five weeks.
RoomMind AI wireframes for onboarding, room selection, photo upload, processing and style choice
RoomMind AI wireframes for onboarding, room selection, photo upload, processing and style choice

The solution

The flow keeps people oriented while the AI works and makes the before and after easy to compare.

Onboarding and taste profile

Three questions come first, starting with what best describes your space, each answer a large card. Then a swipe deck of styled rooms, right to love and left to skip, with style tags and a taste profile bar that fills as you go. A plan screen closes onboarding, with Continue with free plan as visible as Start with Pro. By the first redesign, the app already knows the person's taste.

Home

Home leads with a New Design card and a Start Project button, then recent projects with their style and date, then style inspiration. A floating four icon tab bar keeps the main areas one tap away. Starting something new is the biggest target on the screen, because that is why people open the app.

Room type and photo upload

Step 1 of 3 asks what space you're reimagining, with photo cards for living room, bedroom, kitchen, office and more, each with a short line such as "Quiet sanctuary". The choice shows at the bottom beside Next. The upload screen takes a new photo or one from the gallery, with a tip to shoot a well-lit room and capture as much of it as possible, plus recent photos to reuse.

AI room scan and confirmation

While processing, the photo stays on screen with scan points and a percentage, and a checklist names each step: analysing architecture, identifying light sources, mapping dimensions. The confirmation screen then tags what was captured (architecture, dimension scale, lighting) and notes the room geometry it logged. The wait reads as work being done, and people see the AI understood the room before choosing a style.

Style selector and refinement

Six styles (Minimal, Industrial, Modern, Luxury, Scandinavian, Japandi) are photo cards, not a list of names, because people recognise a style faster than they can name it. The refine screen sets a colour palette, a lighting mood such as bright or moody, furniture density on a slider and a budget range from one to three dollar signs, with a step counter.

Result and regeneration

The result is a before and after slider over the room, tagged AI generated and high fidelity, with the style name and a short description. Below it, Refine Your Vision lets people tap a zone to regenerate just that part, and Reset AI starts over. An error screen handles the moment something goes wrong, and an alert tells people when a redesign is ready.

RoomMind AI screens: room selection, AI processing, a before and after result, style choice and refinement options
RoomMind AI screens: room selection, AI processing, a before and after result, style choice and refinement options

The result

  • The iOS app UI, more than 40 screens: onboarding, home, upload, AI scan, styles, results, plans and error states.
  • The user flow, wireframes, persona and a UI style guide.

How it helps the client

The design takes someone from a photo to a redesign they shaped themselves, so the result feels directed, not random. The client gets onboarding that learns each person's taste first. That's the aim of my product design for AI apps, and there are more examples in my AI case studies.

Frequently asked questions

  • From this project: a short taste quiz and swipeable style cards, room type selection, photo upload with framing tips, a visible AI scan, style and refinement controls, a before and after result with zone by zone regeneration, saved projects, and alerts when a redesign is ready.

  • I show the work. In RoomMind the processing screen lists what the AI is doing (analysing architecture, finding light sources, mapping dimensions), then confirms what it captured before the result. People stay oriented, and the redesign feels like something they directed rather than something that happened to them.

  • Ask before the first result, not after. RoomMind starts with three quick questions about the space, then a swipe deck where people love or skip styled rooms while a taste profile bar fills. Style choice later uses photo cards rather than words, and the refine screen adds colour, lighting mood, furniture density and budget. The first question says answers can be refined later, so nobody gets stuck, yet the first redesign already reflects what the person picked.

  • The photo decides the quality of everything after it, so the upload screen coaches before it asks. In RoomMind people first pick the room type, then take a photo or choose one from the gallery, with a tip to use good light and fit as much of the room in frame as possible. Recent photos sit below for a quick retry. After the scan, a confirmation screen shows what the AI captured, so a bad photo is caught before a redesign, not after.

  • The deck's timeline runs five weeks. Research and market analysis start in week one, UX strategy and AI planning follow, wireframes and UI design run through the middle, and user testing, developer handoff and UI improvements close out weeks four and five.

  • Yes. I build iOS and Android apps in React Native or Flutter from a documented design system, so the build matches the screens. You get a written update every 48 hours and a review call each week, from the first sprint to handoff.

More case studies

Nexfolio investment dashboard design: AI Insight screen with portfolio value, AI confidence, sentiment and a future forecast

Fintech · 2025

Nexfolio

An AI-powered investment dashboard

Lyvio AI real estate app and website design cover: the Lyvio name over a desert dune and terracotta villa

Real estate · 2026

Lyvio

Website and product screens for an AI property advisor

Planning a product like RoomMind AI?

A free 30 minute call. You leave with a clear next step, whether we work together or not.

Book a free intro call