AI Home Remodel
Launched AI Home Remodel from concept to release and beyond. I owned the core experiences: defining a style library to guide AI generation, generating designs, removing unwanted objects from photos, and finding products to bring an idea to life.
AI Home Remodel
An AI tool that turns a photo into design ideas for interiors, exteriors, and landscaping.
Context
Before a renovation, it is hard to picture how a design will look in your own space. Commissioning a visualization takes time and money, and still leaves room for mistakes.
AI Home Remodel lets users upload a photo, choose a style, and visualize a new interior, exterior, or outdoor space.
Discovery
During discovery, I explored the main use cases for AI-powered space design. This informed the structure of the first release and a style library to guide generation.
1. Generating designs
Design generation is the app's core feature. After several iterations, we arrived at a clear flow: upload a photo, select the space type and style, and receive an AI-generated design.
One key improvement was adding a space-type selector before style selection. Interiors, exteriors, and outdoor spaces need different styles and generation settings. This way, users see only relevant options, and the result better matches the space in the original photo.
Generation took 40–50 seconds, so the waiting experience needed special attention. I added a clear processing state and visual feedback so the app would not appear frozen.
2. Removing unwanted objects
The goal was to remove distracting objects from a photo while leaving the rest of the space unchanged.
Objects vary in size, so we added an adjustable selection size. The same tool can handle both small details and larger objects.
3. Finding products
We designed product search as a potential retention driver. Finding similar products from a photo was intended to give users another reason to return when looking for a specific item for their space.
Results
Working to a tight timeline, I designed AI Home Remodel from scratch and prepared it for launch. The first release delivered the core journey from a photo of a space to an AI-generated design.