AI-Powered Search Platform
Semantic search for Keystone Properties' listing inventory, letting renters search by intent instead of rigid filters.
Customer: Keystone Properties (demo)
Read the full case studyKeystone Properties manages residential listings across multiple markets with highly variable descriptions. We built a semantic search layer that lets prospective renters search by intent — describing what they want in plain language — rather than navigating a long list of filter checkboxes.
Challenge
Keystone's existing filter-based search required renters to know exactly which checkboxes corresponded to what they wanted, and listings with inconsistent or incomplete structured data were effectively invisible to filtered search even when they were a good match.
Solution
We built a hybrid search system combining semantic vector search over listing descriptions with traditional structured filtering. Renters can type a natural description of what they're looking for, and the system returns relevant listings even when the exact wording doesn't match — while structured filters remain available for renters who know precisely what they want.
Architecture
Listing descriptions and metadata are embedded into a vector index, updated as listings are added or changed. Search queries are processed through both the vector index (for semantic relevance) and traditional structured filters (for hard constraints like price range or bedroom count), with results merged and ranked. The frontend surfaces both natural-language search and traditional filters side by side rather than forcing a choice between them.
Key features
Natural-language search
Renters describe what they want in plain language instead of selecting filters.
Hybrid ranking
Semantic relevance combined with structured filters for hard constraints.
Real-time index updates
New and changed listings are searchable within minutes.
Fallback to traditional filters
Structured search remains available alongside semantic search.
Integrations
Listing management system
Automatic re-indexing as listings are added, updated, or removed.
Representative outcomes
Improved match relevance
Listings with inconsistent structured data became findable through semantic matching.
Fewer zero-result searches
Natural-language queries surface relevant listings that keyword filtering would miss.
Faster listing discovery
Renters find relevant options without needing to know the right filter combination.
Technologies
Have a similar project in mind?
We're glad to talk through the approach, even before you're ready to commit.