By teaching cameras to understand movement and nuance, Lytehouse reshapes how physical spaces keep people safe

The cameras aren’t just watching; they’re thinking and working. That’s the quiet shift happening inside Singapore’s FairPrice Group stores, where Lytehouse’s software turns video feeds into living systems of awareness.

“Our focus is simple: creating safer, smoother, and more human-centred in-store experiences,” states Natalie Doran, the chief executive officer and co-founder of Lytehouse.

So, instead of rows of monitors flickering with CCTV footage, the system spots what matters: a spill on the floor, a queue forming, or a customer who might need help.

It’s an idea that captures both Doran’s approach to technology and her unconventional path into it. Before founding Lytehouse in 2019, she wasn’t an engineer or computer scientist but a storyteller — one who built entire worlds from imagination.

Rewriting the script

Doran began her career writing animation scripts and developing worlds for children’s games, a role that taught her to think in systems: how each element connects, moves, and evolves. That instinct for storytelling turned out to be unexpectedly helpful in the startup world.

“A storyteller’s job is to innovate, to create new ideas and turn complex narratives into concepts that connect on a human level,” she says. “I didn’t realise it then, but that skillset enabled me to break out of the writer’s mould and drive real-world impact through technologies that can reshape our communities.”

Never going to business school or working for a major tech firm also actually helped. “Startups are by nature a rebellion against the status quo,” she says. “I’ve always been the hungry underdog, pushing myself to the edge of my comfort zone.”

Doran adds: “It’s the grind, the learning, and resilience from over a decade in the startup trenches that gives me an edge. You don’t need cookie-cutter credentials to make a dent in this space.”

It was that same openness to unlikely pairings that led her to Lytehouse’s co-founder, Jean-Vicente De Carvalho, and to the idea that technology could carry empathy as well as precision.

Doran met De Carvalho, now the startup’s chief technology officer, at Entrepreneurs First, where she heard how his family in South Africa had been robbed 37 times at gunpoint. “Growing up in a small town in the UK, violent crime was practically non-existent,” she recalls.

“Although we had different experiences, we bonded over the idea of making any community as safe as Singapore. If we could save (even) one life, it would be worth it.”

That conviction — to build technology that works with people rather than over them — became Lytehouse’s foundation.

The company’s platform applies reasoning, not just recognition, to the physical world, learning to interpret behaviour, anticipate risk and support people running complex environments.

Play

The insight came from a small experiment in De Carvalho’s father’s retail store. “He was sick in the hospital, so we installed our prototype, Scout, our human-intelligence model, to monitor for suspicious activity,” Doran says.

“Lytehouse detected the cashier stealing from the till, and a second individual appeared to be an accomplice. The movement was so subtle we had to watch it three times to see it. That’s when we realised the true power of Lytehouse is in reasoning — not reacting.”

Teaching machines to reason

Traditional video analytics hunt for specific actions or objects, she explains, but “the real world is messy and unpredictable”. Businesses end up relying on human operators to sift through hours of footage. Lytehouse reverses that equation: It trains its intelligence to think like a store manager, drawing context from comprehension, not pixels.

That ability caught the attention of the FairPrice Group, Singapore’s largest retailer, which “operates stores that are living ecosystems with thousands of moving parts”. In its Store of Tomorrow initiative, Lytehouse acts as a digital co-worker: a retail manager, safety officer, and security guard rolled into one.

“It’s a blueprint for how Lytehouse can deliver real value beyond traditional risk management,” she says. “We get to co-create the future of retail spaces and test how video, sensors, and automation can talk to each other to make stores feel as convenient as online shopping.”

The system now automates everything from spill detection to queue management and trip-hazard alerts. Each event becomes part of an actionable report that helps staff prevent incidents, optimise flow, and improve the shopping experience minute by minute.

Some of the most valuable insights, Doran adds, aren’t flashy dashboard metrics but small interventions that change what happens on the floor — the kind of improvements that save time, prevent accidents, and lift customer service.

The partnership has also given Lytehouse a model that can scale to other major retailers. As each deployment adds new data, the system grows sharper, adapting across industries from health and safety to logistics and manufacturing. The company has raised US$1.9 million ($2.48 million) to date.

What matters most to Doran is how this technology interacts with people. “Technology should give us back time — time to build relationships, to innovate, to think, to connect with our environments, families, and customers,” she says. “Artificial intelligence (AI) should handle the repetitive and the routine so we can focus on what makes us human.”

She calls Lytehouse a “co-pilot” for the physical world, one that spots what matters, so humans can act. “No one can stare at hundreds of video streams for eight hours a day,” she says. “In high-risk environments, one missed event can cost lives. Every new camera and user makes Lytehouse smarter, and every new use case supports thousands of workers.”

Her north star, she highlights, isn’t a milestone but a feeling: “Sitting back one day and knowing that the blood, sweat, and tears changed the world for the better. And that I became a leader who inspired others to do the same.”

Precision without overreach

It’s a grounded outlook for someone building intelligence at scale, one that keeps her focused on people even as the systems grow more capable. As AI becomes more embedded in everyday life, Doran is mindful that the same systems designed to help can also overstep. Privacy, she asserts, is non-negotiable.

“Our intelligence is focused on store activity, to identify problems, create shopper insights and increase efficiency — not on the personal information of specific individuals.”

The emphasis is on understanding environments so they run smoothly and safely. Context matters: “Retail environments are tricky, with lots of people, lots of movement. A safety risk in a store versus a mine is totally different. Even between a store in South Africa and one in Singapore, the context changes everything.”

Lytehouse localises its models site by site, blending video data with large-language models so customers can train the system in plain language, “just like they train their staff”.

Doran’s sights are set on scaling globally, but her measure of success remains human. “We deal with high-risk sites, and what’s at stake are lives,” she says. “That pushes us to build better because what we do matters.” 

In five years, she hopes Lytehouse will have turned every space into a connected environment where people are safe, engaged and supported by intelligence that feels invisible. “This is bigger than CCTV,” she says. “It’s about making our environments intelligent enough to serve us. I want people to see a camera, think Lytehouse, and know we have their back.”