Blinko
All case studies
Case study · Real estate agency · Sydney, Australia

AI matching built to save a Sydney real estate agency an estimated 2,400 hours a year

We built an AI system that cleans and matches property records automatically, so agents spend their time on clients instead of spreadsheets.

hours a year, estimated at handover (about 1.2 full-time roles)
~2,400
hours a year, estimated at handover (about 1.2 full-time roles)
hours per agent each month, estimated at handover
~10–12
hours per agent each month, estimated at handover
of match confidence, so agents know where to start
3 tiers
of match confidence, so agents know where to start
What changed

Before and after

  • Before: Agents matched property records to listings by hand

    After: AI cleans and matches them automatically

  • Before: Nobody knew which matches to trust

    After: Every match is rated and explained

  • Before: Slow, uncertain matching delayed outreach

    After: Matches are ranked, so agents start with the strongest

The problem

Agents had to match the agency's property records with outside listings, and the data never lined up. They did it by hand for hours every month, and still did not trust the results enough to act quickly.

What we built

An AI matching system connected to the agency's CRM. It cleans incoming records, finds the matches, and tells agents how confident it is in each one.

  • AI data cleaning
  • AI matching
  • Confidence ratings
  • CRM integration
Every match rated high, medium, or low confidence.
Every match rated high, medium, or low confidence.
How it works

How the matching works

AI handles 3 of 6 steps. People make the final call.

  1. 1
    Automation

    Collect

    Pulls in property records from the agency's systems and outside listings.

  2. 2
    AI (LLM)

    Clean

    AI tidies up addresses, names, and descriptions so records can be compared.

  3. 3
    AI (LLM)

    Match

    AI finds records about the same property, even when they are written differently.

  4. 4
    Scoring

    Rate

    Each match gets a confidence rating: high, medium, or low.

  5. 5
    AI (LLM)

    Explain

    AI writes a one-line reason for each match, so agents can trust it.

  6. 6
    Your team

    Act

    Agents contact the best matches first, and everything is logged in the CRM.

The results

The bottom line

At handover we estimated the system would save each agent 10 to 12 hours a month, about 2,400 hours a year across the agency. The client called it “incredibly valuable”.

Next case study

Construction fintech

Fractal runs $100M+ a year in transactions on one live system we built

Read it

Want results like these?

In a 30-minute call we look at the process you want to fix and tell you honestly what it would take.