Built BI Match in one day, from idea to stakeholder decision, and shipped it to production

Browsing a list isn’t how people choose a home

Buyers know what life they want before they know which apartment, yet the site asked them to scroll a list. The growth hypothesis: matching buyers to residential complexes by lifestyle, through swiping, could convert better than a traditional listing.

BI Match: find a home that fits your lifestyle

BI Match turns search into three steps. Buyers set the parameters of their dream apartment, swipe through the floor plans they like, then leave their contacts to get a personal offer from sales.

• Step 1: choose the parameters of your dream apartment
• Step 2: swipe the floor plans you like
• Step 3: leave your contacts and get a unique offer

Prototype at production fidelity, not in Figma

I designed the interaction model and flow, then built the prototype in our real tech stack with Claude Code, using the same API, component library and data as production. It was functional enough to hand off through GitHub, not a throwaway demo.

• Reframed listing browsing as swipe-based matching
• Designed the full interaction model and user flow
• Built it in the team’s production stack
• Handed off working code through GitHub

From a one-day prototype to a live product

Stakeholders used a working product the same day the idea came up. The team took it forward, and BI Match is now live in production. The point isn’t the AI tooling. A hypothesis that would normally wait weeks for engineering was tested in hours without pulling anyone off their work.