PropTech · USA
Mirabilis AI Match
- Client
- Mirabilis
- Year
- 2025
- Services
- AI agents · Product design · UX/UI · RAG
Overview
Home search fails when buyers type filters they do not fully understand. For Mirabilis we designed an AI Property Match experience: buyers describe lifestyle and constraints in plain language, and the agent returns ranked homes with transparent match scores and “why this fit” reasons — so sales conversations start from intent, not a spreadsheet of listings.
The challenge
Generic listing search is keyword-blind. The product needed grounded ranking over inventory and buyer context — without inventing homes or hiding why a match scored high.
Our approach
We prototyped a retrieval + ranking agent over listing attributes, location signals and stated preferences, wrapped in Mirabilis’s warm white/orange product UI. Natural-language queries map to structured constraints; each card exposes a match score and explainable factors. In the scoped demo, top-3 relevance felt usable for first outreach and cut “filter thrash” for buyers who only knew the lifestyle they wanted.
- AI
- PropTech
- Matching
- Product
Have a project like this?