From design to data: how buildings are learning to adapt

The world didn’t just change in 2020, it accelerated. Work shifted, cities paused, and suddenly buildings around the globe were operating on outdated rules. Office towers sat half-empty. Energy costs soared. Entire floorplans became irrelevant. And through it all, most buildings were still flying blind.

“We’ve been operating billion-dollar buildings based on guesswork,” says Chang Liu, a San Francisco–based spatial intelligence analyst. “In 2025, that just doesn’t cut it anymore.”

Liu is part of a rising movement that believes buildings shouldn’t just contain life, they should understand it. His work sits at the intersection of architecture, data science, and ethics, in a field that’s becoming more critical with every passing year. 

As AI filters into the physical world, someone has to ask: What if buildings could learn? What if they could adapt? What if they could… care?

“We need to think about our interaction with the biggest interface in our daily lives,” Liu says. “How can you have a ‘conversation’ with the space you in, and is it possible the building and the people achieve a simple goal together?”

Liu didn’t start in AI, or data, or even tech. He trained as an architect, then as an urban designer. At Carnegie Mellon University, he spent long hours studying how people move through public spaces — where they pause, how they gather, what flows and what doesn’t. With this in mind, he began stepping into data, interaction, and eventually AI, not to replace architecture, but to deepen it.

“I was always looking for how the building talks,” he says. “Not literally — but whether it’s telling the truth about its purpose. Whether people are using it the way it was meant to be used.”

Today, Liu works at Butlr, a California-based company pioneering privacy-first people sensing. Unlike cameras or facial recognition, Butlr’s thermal sensors capture only heat signatures and motion. No visuals. No identity. Just presence, and patterns.

“For me, privacy isn’t optional,” Liu says. “If people feel watched, they behave differently. And if you’re trying to understand honest behavior, that ruins the data before you even begin.”

At Butlr, Liu helped build the company’s spatial intelligence function from scratch, designing tools that now power smart buildings, corporate campuses, senior care homes, and more.

Ask Liu what his work actually does, and the answer is refreshingly clear: 

“We help buildings understand what’s really happening inside.”

That understanding shows up in different ways for different clients. For one university, Liu’s team mapped thousands of hours of classroom data — revealing that nearly 40% of scheduled lecture halls were chronically underused. For a Fortune 500 company, his analysis showed that high-demand collaboration spaces were routinely maxed out midweek, while other zones sat empty. The insight helped them rebalance their layout, reduce complaints, and delay costly construction.

In another project, his team flagged entire floors where HVAC systems were running at full power for rooms with near-zero occupancy. Over the course of a year, the wasted energy was equivalent to hundreds of thousands of dollars — and tons of unnecessary CO₂ emissions.

Liu’s work gives real estate teams a clear picture of what’s actually happening inside their buildings. Instead of relying on assumptions, they can redesign workspaces around real behavior, fine-tune cleaning and maintenance schedules based on demand, and cut HVAC waste by tying airflow directly to live occupancy. The same intelligence lets them simulate future space needs before committing to costly construction, turning guesswork into evidence-based planning.

One of his favorite use cases? Catching ghost meetings, calendar-booked rooms that are never actually used. “We’ve seen buildings where up to 70% of reserved rooms are empty,” he says. “That’s a huge opportunity for many many offices, you don’t need more rooms, you just need smarter scheduling.”

Because the data is real-time, privacy-safe, and spatially precise, it unlocks a new class of dynamic building systems. HVAC can adjust minute by minute based on how people actually move. Room-booking platforms can reflect real behavior instead of calendar fiction. And workplace layouts can evolve continuously as patterns shift, allowing buildings to respond as fluidly as the people inside them.

“It’s not about surveillance,” Liu emphasizes. “It’s about clarity. About giving space the ability to respond intelligently, just like any other part of a modern system.”

Across campuses, offices, hospitals, and even museums, Liu’s work is helping shift the paradigm: away from static square footage metrics, and toward behavior-driven intelligence.

But the work doesn’t stop at buildings. For Liu, the most powerful use case is much more human: senior care.

“It’s personal,” he says. “These aren’t just data points, they’re our parents, our grandparents. If we can use space to keep them safer, more comfortable, that’s worth everything.”

Liu’s team has deployed their system in senior-care facilities across the U.S. and Asia, helping caregivers spot subtle behavioral shifts long before they become medical emergencies. One facility noticed a slow but steady rise in bathroom dwell time, a signal of early mobility decline. The changes weren’t visible to staff, but the system picked them up weeks in advance.

“That’s the magic,” Liu explains. “You don’t need to watch people to see what’s changing. You just need to understand the rhythm of the space.”

In another case, aggregated heat signatures revealed nighttime restlessness in a memory care wing. This insight led staff to adjust sleep routines and reduce fall risks. On yet another note, the team detected unusually long durations in a specific restroom, prompting a facility-wide redesign that improved accessibility for residents with slower movement patterns.

By combining anonymous presence data with spatial analysis, Liu’s tools give caregivers a new layer of insight: early warnings for decline, invisible patterns of risk, and even clues about how building layout affects resident behavior.

Importantly, all of this happens without cameras or invasive monitoring. The sensors capture only motion and dwell time, no faces, no identity, just behavioral signals.

“This isn’t about surveillance,” Liu says. “It’s about dignity. And it’s about catching the things we used to miss.”

The data also helps facilities benchmark themselves against others, providing operators with new ways to improve care, allocate staff more efficiently, and validate decisions with real evidence.

What began as a smart-building tool has quietly become a health-and-safety engine — one that enhances clinical outcomes, supports overworked staff, and restores agency to aging populations.

To Liu, what’s happening now is just the beginning.

As AI systems grow more capable, and more embedded in the real world, buildings will need to become more than static assets. They’ll need to sense, simulate, and adapt. Liu’s team is already building multi-agent simulations that let buildings anticipate behavior in real time.

“Eventually, there won’t be ‘building tech’ or ‘IoT’ or ‘smart cities,’” he says. “It’ll just be physical-world intelligence. A new layer that sits between people and space, and helps both sides understand each other.”

He describes a future where offices quietly reshape themselves based on true demand, airports redirect flows to reduce wait times and stress, hospitals adjust cleaning and staffing according to actual behavior rather than rigid schedules, and senior-care homes monitor subtle changes in pace, posture, or sleep to provide safer and more dignified support. “Eventually, there won’t be ‘building tech’ or ‘IoT’ or ‘smart cities,’” he says. “It’ll just be physical-world intelligence — a new layer between people and space that helps both sides understand each other.” 

In Liu’s vision, the building becomes part of the team: it listens, it learns, and eventually, it helps.

Despite the scale of his impact, Liu doesn’t posture or self-promote. His influence comes not from grandstanding, but from building things that work.

“I’m not chasing titles,” he says. “I lead through action, by shaping tools, insights, and systems that help people and spaces work better together.”

But his influence is growing:

• His insights have shaped office redesigns across the U.S. and Europe.
• His work has helped senior-care teams catch unseen risks.
• His tools are used by major real estate firms, design studios, and facilities teams around the world.
• He’s part of the award-winning team behind Butlr, recognized by Core77, Good Design, K-Design, and Dezeen.

He blends design intuition with statistical rigor and ethical clarity — a rare mix in an industry trying to figure out what the future looks like.

“Once you see the patterns,” he says, “you can’t unsee them. And once buildings start seeing, too — that’s when the real change begins.”

This article was written in cooperation with in cooperation with Tom White