Project Overview
Syversten is a hybrid real estate platform that combines apartment browsing with AI-powered interior visualization. Buyers can explore available properties and instantly preview how each space could look after renovation — different layouts, finishes, furniture, and lighting — without leaving the listing page. The product was built for a Norwegian client who wanted to bridge the gap between the “cold” property search experience and the emotional decision of imagining yourself living in a place.
- Duration 3-4 months
- Location Remote (client based in Norway)
- Domain PropTech / Real Estate
How we work
We started with a discovery phase focused on understanding two very different user mental models living inside one product — the casual browser scrolling through listings, and the engaged buyer experimenting with renovation ideas. Our UX team mapped both flows separately before designing the transitions between them, while our frontend developers prototyped the AI visualization layer in parallel to validate technical feasibility early. The team operated in weekly sprints with direct client communication, iterating on designs based on real usage scenarios rather than assumptions.
Task
Most real estate platforms stop at the listing — photos, floor plan, contact button. But buyers don't make decisions from photos; they make decisions from how a space feels once they imagine living in it. The challenge was to design a single interface that handled both jobs without becoming overwhelming: fast property search for users still exploring, and a creative AI-driven visualization tool for users ready to project themselves into a specific apartment.
Result
We delivered a complete UX/UI system and frontend implementation built around two seamlessly connected modes. The browsing experience stays clean and conventional — users can filter, compare, and shortlist apartments without friction. Once a user opens a listing, the AI visualization layer activates: they can swap finishes, generate furnished variants, and adjust lighting in real time. Around 20 screens were designed and shipped, covering the full journey from search to visualization to inquiry.
Ready to Build Your PropTech Platform?
Landing Page Overview
The landing page introduces Syversten’s hybrid approach within the first scroll — a clear value proposition, a short demo of the AI visualization in action, and a transition into the property browsing flow. We avoided the typical PropTech cliché of generic city skylines and instead used real apartment renders generated through the platform itself, reinforcing the product’s core promise.
App Overview
Inside the app, the experience is organized around two intentional modes. The browsing mode behaves like a familiar real estate marketplace — filters, map view, saved properties. The visualization mode opens directly from any listing and lets users experiment with the space itself: different interior styles, furniture configurations, and renovation scenarios powered by AI. The transition between modes is designed to feel like a natural next step, not a separate tool, which is what makes the hybrid approach actually work for buyers rather than just sounding good in a pitch.
AI Integration as a UX Decision
The biggest UX challenge in any AI-powered product isn’t the AI itself — it’s helping users trust it without turning the process into a black box. For Syversten, we deliberately designed the AI as a visible tool, not magic happening behind the curtain. Users can see which inputs the system is using — room dimensions, style preferences, material choices — and adjust any suggestion directly. This transparency mattered more than any algorithmic optimization: it turns a skeptical buyer into an active collaborator who shapes the result, rather than passively accepting it.
Two Audiences, One Interface
Syversten had to work for two very different user groups — first-time buyers visualizing an apartment for the first time, and professional interior designers running through multiple variants quickly. Instead of building two separate modes or overloading the interface with every possible option, we used progressive disclosure: simple tools are immediately visible, advanced features appear only when the user signals they’re needed. Beginners are never overwhelmed, and professionals are never slowed down by training-wheel UI.
Performance as a Design Constraint
Real-time visualization in the browser sets hard limits — anything over a two-second delay breaks the feeling that you’re actually interacting with the space. Performance wasn’t a technical side-concern in this project; it was a design decision running through every screen. We reduced the number of options visible at once, optimized asset loading based on what the user was actually looking at, and designed transitions that shorten perceived wait time rather than trying to mask it. Speed became part of the visual language, not just a technical metric.
