1. What Abora is
Property search is solved. Every portal carries the same listings, and finding out what is for sale takes minutes. Property research is not solved, and it is the part that decides whether a purchase was a good one.
Abora sits after the search. Give it an address and it resolves that address to its parcel on the state property map, measures the block from the title geometry, and works out what governs it — easements, planning overlays, orientation, school zones, what comparable homes have actually sold for. Give it a suburb and it builds the same picture at that scale from the Census and recorded sales. Give it only a budget and a brief and it ranks what fits against the things you said mattered.
Behind that sits a fair amount of machinery — spatial analysis across millions of parcels, and AI agents that query the underlying data rather than recall it. How all of it works is written down in methodology and how we score, in more detail than is commercially comfortable, which is rather the point.
Institutional-grade property data, hidden suburb insights, and the working shown.
2. The third generation of proptech
Australian property technology has moved through two clear generations, and is part-way into a third.
First generation
Listings went online
Classifieds became searchable portals. The problem being solved was distribution — getting every property in front of every buyer. It worked, and it is finished. Every portal now carries substantially the same inventory.
Second generation
Data got aggregated
Property data platforms assembled sales histories and demographics, and automated valuation models turned them into estimates. Real progress — but the output was a single number per property, and the reasoning behind it stayed inside the vendor.
Third generation
Analysis became interrogable
Spatial databases can now hold millions of parcels and answer geometric questions in milliseconds. State governments publish cadastral and planning data openly. And AI agents can query primary sources on demand instead of working from a pre-computed summary. This is the layer Abora is built in.
The shift that matters technically is from pre-computed to interrogable. A second-generation platform decides in advance what it will tell you about a property and stores the answer. Abora holds the primary data — parcel geometry, planning overlays, easements, school zones, Census tables, sale records — and runs the analysis against your question when you ask it. That is why the ranking in a search can be rebuilt around one buyer’s priorities rather than served from a generic index.
What the AI is actually doing
“AI-powered” is doing a lot of unearned work across this industry, so it is worth being precise about ours. Abora runs a multi-stage agent architecture: large language models with tool-calling access to the underlying property and demographic databases, orchestrated as state machines rather than single prompts, with a governed registry controlling which data capabilities each agent can reach.
A property search executes as roughly seventeen coordinated stages — normalising the brief, filtering candidates, enriching each with geospatial context, scoring multiple dimensions in parallel, then aggregating and explaining — and streams results as each stage completes. The natural-language tools work the same way: the agent selects a tool, the tool runs a real query, and the answer is built from what came back.
The distinction we hold to: machine learning and language models organise and explain the analysis; they never produce the measurements. Every figure — land area, easement coverage, travel time, median price, match score — comes from deterministic spatial computation. An AI that must call a tool to get a number cannot invent one, and that constraint is what makes the rest of it trustworthy.
3. Why it works this way
The industry standard is to take a great deal of data, compress it into a single confident number, and show you the number. It looks authoritative and it is easy to act on. It is also where the useful information goes to die — the moment a score exists, the facts that would have complicated it stop being shown.
We think that trade is backwards for someone spending most of what they have. What a buyer needs is not more confidence, it is more legibility: the fact, what it enables, what it constrains, how the neighbours compare, and where the figure came from. That takes longer to read than a score out of ten. It is also the only version you can check, argue with, or take to a conveyancer.
So the judgement stays with you. Abora sets out what the data shows — the fact, what it enables, what it constrains, how the neighbours compare — and leaves the decision where it belongs. That takes longer to read than a verdict. It is also the version you can check.
4. How we handle the data
These are design constraints in the product, not sentiments.
- Every number carries its source and its date. If we cannot say where a figure came from, it does not go on the page.
- Absence is stated, not hidden. When there are too few sales for a median to mean anything, we say so and say why, rather than showing a number built on two transactions.
- Numbers are computed, never generated. Measurements come from spatial queries and arithmetic. AI organises and explains what was measured; it is never asked to produce a figure.
- We say what we do not cover. Property X-Ray and Suburb Insights are Victoria-only, because the state data behind them is. We would rather say that than serve a thin page for a suburb we do not know.
- We publish places, not people. Our data describes land and localities. We do not collect or publish owner or occupant information.
5. Limitations
We do not give financial, legal or planning advice. We hold no Australian Financial Services Licence and we are not licensed valuers. On anything that decides a purchase, verify with the council, a conveyancer, or a licensed valuer. See our terms.
6. Where we are
Abora is built and run in Melbourne by Abora Pty Ltd, an Australian company.
Property X-Ray and Suburb Insights cover Victoria, which is where the cadastral and planning data they depend on is published. Market Pulse and the calculators are national. Where a tool is limited to one state, it says so at the point you use it rather than in the fine print.
7. Get in touch
If a figure here disagrees with its source, that is a bug on our side and we want to know — tell us and we will tell you what our source says. For partnerships, media or anything else, hello@abora.com.au.
If you want to see how the work is done before you trust it, start with methodology — it lists every dataset behind the site, and what happens between the source and the page.