We use cookies to improve your experience.

    Back to case studies
    PropTech case study: after 6 months in production: with SelectCursor
    PropTech

    After 6 months in production:

    40% Reduction in manual matching time10K+ Properties on platform3mo From idea to production MVP

    The Challenge

    Propty, an Amsterdam-based PropTech startup, identified a critical pain point in the Dutch rental market: property managers were spending hours manually matching tenants with suitable properties. The process was inefficient, error-prone, and created bottlenecks that frustrated both property managers and prospective tenants.

    The key challenges were:

    Manual matching overload โ€” Property managers reviewing 50+ applications per property, taking 2-3 hours per listing

    Subjective decision-making โ€” No standardized criteria led to inconsistent tenant selection

    Missed opportunities โ€” Good tenants fell through cracks due to delayed responses

    Our Approach

    We partnered with Propty to build an AI-powered property matching platform that would automate tenant-property matching while maintaining transparency and compliance.

    We started by deeply understanding the problem:

    Key insight: The matching problem wasn't just about filtering โ€” it was about prioritization . Property managers needed help identifying the best matches quickly, not just eliminating poor fits.

    We designed a multi-component AI system:

    We built the platform using our proven stack:

    The Outcome

    After 6 months in production:

    40% reduction in time spent on tenant-property matching

    25% improvement in tenant retention (better initial matches)

    3x faster time-to-fill for vacant properties

    10,000+ properties now on the platform

    โ€œ"Selectcursor didn't just build us software โ€” they built us a competitive advantage. Our property managers now focus on relationships instead of paperwork. The AI matching has become our secret weapon in a crowded market."โ€
    CEO, Propty,

    Technologies Used

    React
    OpenAI
    AWS
    AI
    PostgreSQL
    pgvector
    FastAPI
    Python

    Timeline

    • We started by deeply understanding the problem:Phase 1: Discovery & Validation (2 weeks)
    • We designed a multi-component AI system:Phase 2: AI Architecture Design (2 weeks)
    • We built the platform using our proven stack:Phase 3: Development & Training (6 weeks)
    • We launched with a controlled rollout:Phase 4: Launch & Iteration (4 weeks)

    Building something similar?

    Book a call to discuss your project.

    [email protected]We respond within 24 hours