Buying a home is one of the biggest decisions people make, but the way we search for them hasn’t kept up.
It still looks the same: set a few filters, scroll through listings, and try to match what you see with what you want. But how a home feels doesn’t fit neatly into filters, and a lot gets lost in the process.
At the same time, more buyers are starting somewhere else entirely: in AI.
That shift exposes a deeper problem; while the way people search is changing, the underlying data isn’t. It’s fragmented, shallow, and not built for machines to understand, which is why the experience breaks down the moment you move beyond filters.
The system was never designed for how people actually search
Real estate platforms were built for a different era, when search meant selecting from predefined options and working within their constraints.
But that’s not how people think about homes. People communicate housing preferences through subjective cues rather than structured criteria: a bright kitchen, a place that feels modern but still lived-in, somewhere close to a good school but not on a busy street. These are not clean inputs. They are subjective, layered, and often hard to articulate precisely.
Current systems simply aren’t designed to handle that. Those signals get lost because the data underneath doesn’t capture them in a usable way.
This gap is becoming more visible as AI becomes the starting point for discovery. When people begin searching in natural language, the limitations of existing data structures are no longer hidden behind filters, they become the bottleneck.
Building the layer AI actually needs
Neuralindex is building the data index behind every AI answer about property.
At its core, the product transforms messy real estate data into a structure that machines can properly understand. Instead of relying only on structured fields like bedrooms or price, it analyses listing photos, descriptions and location data to extract the kinds of signals people actually care about: whether a space feels bright, whether the home has a good setup for elderly parents with mobility issues, or whether a property has renovation potential.
These signals are then organised in a way that allows AI systems to reason over them, rather than simply retrieve them.
On the user side, this enables a different kind of search experience. Someone can describe what they want in natural language, and the system can interpret that intent and return properties that genuinely match it. The goal isn’t to improve filters, but to move beyond them entirely and align the system with how people already think.
Three perspectives, one problem
Neuralindex didn’t start in real estate. It started in a coworking space in Da Nang, Vietnam where the three founders were working on completely different things.
Naveed Nekoo was working on an AI controlled quadcopter swarm, Mat Heywood was experimenting with LiDAR scans and Son Duong was deep in machine learning. They weren’t collaborating, but they kept circling the same question: if AI can understand physical environments, why can’t it understand homes?
That question led to a broader realisation: property search hadn’t evolved in decades, but the way people interact with information has.
Once they saw that, the direction was clear. Search would move into AI, but the data layer to support it didn’t exist.
The team was built for that gap. One founder brings deep real estate experience and a clear view of how broken the system is for buyers. The other two bring the technical depth to rebuild it.
From concept to real-world use
Neuralindex has moved from idea to product quickly. The team has already built an early version of the system that enables natural language property search in a way that works on real inventory, not just in theory.
They are currently running pilots with agency groups across multiple countries, indexing tens of thousands of listings and deploying the product across hundreds of offices.
That progress has been important in validating the approach, but the more telling signal has been how the product is received. When experienced operators in the property industry see it and immediately understand the shift, it becomes clear that this is not just a marginal improvement; it represents a different way of approaching search altogether.
What property search looks like next
In the short term, the focus is on scaling: expanding the dataset, increasing coverage across markets, and continuing to improve how AI systems interact with real estate data.
Over a longer horizon, the change is more fundamental. Property search is likely to move away from filters entirely, toward systems that can interpret intent directly. Instead of adapting their thinking to fit the platform, users will expect the platform to understand them.
In that world, the most valuable layer is not the interface, but the infrastructure that makes that understanding possible.
Neuralindex is building toward that position, aiming to become the data layer that connects how people describe homes with how machines interpret them.
The ask
Neuralindex is looking to connect with others building in adjacent spaces, particularly in search and AI infrastructure.
As property search shifts toward AI-native experiences, the ecosystem around it is still taking shape, and the team is keen to collaborate with others working on similar problems.
Watch them pitch at Demo Day
When: Thursday, 30 April @ 7:00 PM (pre-party starting at 5 PM)
Where: Carriageworks (at the close of Blackbird's Sunrise Festival)
What: Pitch night (19 companies)
Tickets: Grab your ticket here
