For two decades, finding somewhere to rent in Britain has meant typing a postcode into Rightmove or Zoopla and scrolling. Steven Charlton thinks that is a thin idea of search, and he has built a platform to prove it. nHabit, which reaches the App Store and Google Play in mid-June, lets renters describe the life they want in plain language and hands back neighbourhoods they would never have thought to type in. “Instead of following the herd to the same old postcodes,” Charlton says, “our aim is to let people search the way they’d plan their perfect holiday with ChatGPT.” The idea has already pulled build-to-rent operators including Quintain and Grainger into conversation, and it arrives as Rightmove, which by its own research takes around 80 per cent of the time British consumers spend on property portals, defends a £1.5 billion class action over the fees it charges agents. Charlton, a former Perkins&Will managing director turned founder, used a wide-ranging PropCast appearance to set out why he believes the two incumbents are too big to fix the thing renters actually struggle with.

How the search actually works

The starting insight is almost embarrassingly simple. “You need to know where you want to live before you can search,” Charlton says, “and how can you know all the areas you could live in when you’ve never been to them all?” nHabit flips that around. A renter draws a boundary by travel time, a method Charlton calls isochrone generation, then tightens it with the things that actually shape a day, a ten-minute walk from a Tube station, good schools nearby, and the app surfaces homes in places the renter had never weighed up.

Behind the conversational front end sits Milo, a proprietary large language model wired to a three-dimensional graph database. It is deliberately closed, working only across the roughly 100 datasets nHabit has ingested rather than crawling the open internet, and it answers in whatever language the question is asked. Renters tune five dials, safety, nightlife, amenities, digital connectivity and mobility, to their own priorities. “Everybody’s different,” Charlton says, recalling a South Korean renter who put safety first and still ended up somewhere that felt unsafe for want of the data to choose well, against a group of Australians who cared about nightlife and nothing else.

Taking on the incumbents

The duopoly has barely moved in twenty years, and Charlton is blunt about why. The portals, he argues, cannot rebuild themselves around AI without tearing up the systems they already run on. “This is a ground-up build, not a ChatGPT chatbot wrapper dropped on top of an existing system,” he says, and the two giants are, in his view, simply too big to attempt it. He is just as withering about the wave of look-alike tools claiming an AI edge. “I look at a lot of businesses and think, that’s basically an AI wrapper,” he says. “You’re just piggybacking on somebody else’s technology. It’s essentially a dashboard.” OnTheMarket and others have tried to break the lock before and offered, in his words, alternative versions of the same thing.

The £1.5 billion claim against Rightmove, led by former Competition and Markets Authority panel member Jeremy Newman and funded by litigation specialist Innsworth Capital, reaches its certification hearing in November, and Charlton reads it as a market finally losing patience. He has heard the standard objection plenty of times. One national agency told him he would need venture capital, private equity and a £20 million annual marketing budget to land a punch. “Social media has genuinely levelled that playing field,” he counters, pointing to the direct-to-consumer brands that scaled through COVID on a fraction of the old launch cost. “If the industry is genuinely sick of the status quo, people need to actually support an alternative rather than just complain.”

What it means for landlords, operators and agents

For the operators and agents who pay to be seen, the first benefit is cleaner demand. Matching renters to homes on lifestyle and neighbourhood fit produces better-qualified leads and fewer dead-end enquiries, the difference between a showcase and a switchboard. The deeper prize is the data underneath. “Data is the new oil,” Charlton says. nHabit builds anonymised personas from how people behave in the app, whether they own a dog, what they linger on, what they swipe away, and reads the patterns the way Netflix reads viewing. “Why is it that people with dogs are less age-sensitive than people with children?” he asks. “The data might tell us.” Ownership is the point he keeps pressing: with the incumbent platforms, the insight ends up in someone else’s hands.

That rewrites the commercial model. Rather than a monthly listing fee, nHabit offers a developer the news that a particular profile of renter was ignoring a location six months ago and is now circling it. Quintain, the Wembley Park operator, grasped the idea at once, Charlton says, and a conversation with the build-to-rent landlord Grainger surfaced something he had missed, that many of its tenants work in the NHS on shift patterns whose travel times look nothing like a nine-to-five. The same logic carries into student housing, where operators such as IQ and Unite hold safety credentials Charlton thinks they undersell, and the recently enacted Renters’ Rights Act only sharpens the appetite for better data.

Agents, meanwhile, get a read on roughly twenty renter typologies and on exactly what a prospective tenant is hesitating over. A structural shift sits behind the sell. A year ago, Charlton estimates, about one per cent of an agency’s leads came through tools like ChatGPT or Claude, and he now puts it at seven or eight per cent, noting that those systems crawl websites selectively. “Agents understand they need to get their websites ready for LLMs,” he says, “and we’ve essentially done a lot of that work for them already.” The value spreads wider still, with one of his non-executive directors, who previously led Microsoft’s digital-cities work, pointing to retailers, hospitality and councils as buyers of the same locational insight.

From architecture to a blended business

The route here ran through the top of global architecture. Charlton trained in interior design at Edinburgh College of Art, decided early that he “would rather employ great designers so we could elevate together”, and moved to Dubai to set up the Middle East studio of Pringle Brandon, the commercial-interiors firm the architect Jack Pringle founded with Chris Brandon in 1986. When Pringle, now chair of the RIBA board of trustees, sold the business to Perkins&Will in 2012, Charlton’s remit widened from fitting out offices to winning architecture for the Dubai developers Emaar and Aldar. A country-scale masterplan changed how he saw the work. “It’s about data, understanding what the infrastructure is going to be in ten, twenty, thirty years,” he says, and the conversations that followed, with Siemens and Schneider, planted the idea behind everything since. He became Perkins&Will’s UK managing director in 2017, ran the London studio for four and a half years and left in 2022 to start i/o atelier, named for the binary of input-output and the atelier, a house of artisans, a “blended business” where machine-learning engineers and designers sit side by side.

nHabit emerged from that studio almost by accident. Asked to measure the “vibrancy” of the places i/o was designing, the team-built mapping software that scored London neighbourhoods on amenities, green space, transport, gyms and the rest of the texture of daily life, then triangulated those points of interest into a picture of how appealing an area really was. Bolted onto the language model the studio had already built, the engine turned out to do something else entirely, helping people find somewhere to live. ChatGPT launched roughly six months after Charlton founded the studio, and the doubters came round. “Many people clearly went away and thought he’s lost the plot,” he says. “Over time those same people have come back and said, actually, you were just ahead of the game.”

Ambition, and what has been built

The proprietary work, Milo, the isochrone engine and the graph infrastructure beneath them, is where Charlton sees defensible intellectual property and the prospect of patents. Because the infrastructure is built, new markets switch on quickly: Manchester and Liverpool are ready, and the longer horizon is Paris and New York, cities restless enough to reward the model. “Why can’t we be the Airbnb of residential rental?” he asks, noting that most rental apps he meets abroad are stuck serving a single city. Airbnb itself started in San Francisco.

nHabit is self-funded and was founded only in April 2025, and the design business is heading the same way, toward helping occupiers procure design rather than only producing it. An adviser put the trajectory back to him, that he is “becoming a tech business that does design rather than a design business that does tech”, a verdict offered with equal parts admiration and unease. The name says as much. nHabit is “inhabit” with the i taken out, a small act of rebuilding from the letters up, which is roughly what its founder has set out to do to the way Britain looks for somewhere to rent.