This is our AI data pipeline. AI does the cleansing, the name matching, the corrections and the address normalising — day and night — so ChatGPT, Claude and Gemini Enterprise always read data that's clean, current and ready to use.
The nz.house data engine. Raw records go in the front. Clean, labeled records come out the back.
AI runs every step, on every record, around the clock.
Council files, titles, sales, listings and market news — from every source we can get.
AI matches names and addresses, pinning every scrap to the right home — whatever the spelling.
AI corrects errors, fills gaps and removes doubles — checked before it's served.
AI adds plain-language labels for prospecting and marketing.
New sales and listings flow in all the time. Answers are never stale.
Matching names, normalising addresses, correcting errors — the fiddly work no human could do across two million records. Here's what that looks like:
Raw data tells you what a home is. Labels tell you what to do with it. We label every record two ways:
Signals that point to your next vendor, built from ownership, sales and listing history — so your AI can build call lists, not just look things up.
The angles that sell a home, drawn from the property record and the area — so your AI writes ads, posts and letters that fit the buyer.
Connect nz.house to ChatGPT, Claude, Gemini Enterprise or any AI agent that supports MCP — it takes about two minutes.
See the setup guides ask@nz.house