The current wave of AI lives in documents, emails, and dashboards. The next wave will live in space. Embodied AI means software plus sensors plus motors operating in the real world: machines that can see a corridor, decide what to do, and act on it without waiting for a human. Once AI can move, buildings stop being passive data sources and start being operational platforms. That shift will rewrite how we design, contract, insure, and run real estate.
Why now? Three forces are converging. First, vision data is exploding. You can’t train reliable autonomy on generic images, you need recent footage from actual lobbies, meeting rooms, kitchens, roofs, and plant rooms. Second, simulation has grown up. Digital twins are more than just showpieces, they’re training grounds where robots practice millions of tasks safely before touching your site. Third, utility economics are kicking in. The tech stack that turns photons into decisions into verified actions will behave like a network service. Whoever standardises and operates it across many buildings gains compounding advantages.
So what does this mean for smart buildings?
- Buildings will be staffed by fleets, not shifts. Cleaning, stock moves, deliveries, basic maintenance checks, perimeter patrols. Expect mixed teams of people and robots, dispatched by an orchestration layer that allocates tasks minute by minute and proves they were done.
- The competitive edge shifts to data rights and safety. Owning (and governing) the visual and operational data generated on your floors will determine who learns fastest. Leases and MSAs must spell out who can train models on that data, how long it’s kept, and how it’s anonymised. Safety governance moves from “permits to work” to “permits to operate” for autonomous systems.
- Design becomes robotics-aware. Door widths, ramp angles, thresholds, lift signalling, lighting, floor reflectivity, charging berths, RF coverage, edge compute closets, etc. Many assets carry an “autonomy tax” because they block line-of-sight sensors or trap wheels. New refurb standards will fix that.
- Digital twins get promoted to pre-production. Any change to layout, vendor route, or operational policy gets tested in simulation first. If it fails in sim, it doesn’t go live. Verification and audit come from the same stack, turning compliance from paperwork into telemetry.
- Energy becomes a scheduling constraint, not a footnote. Embodied AI runs on electrons. Duty cycles will align with tariffs, on-site generation, and storage. A clean kWh becomes both cost advantage and ESG proof.
What should owners and operators do now?
- Instrument for sight. Deploy privacy-preserving vision at key workflow nodes. Build labelling pipelines early – they’re slower to stand up than the cameras.
- Set service-level objectives. “95% of lobby debris cleared within 3 minutes”. “Fault triage within 15 minutes”. These become the training targets and the billing basis.
- Create a site autonomy board. Geofences, exception playbooks, incident reporting, insurance mappings. Treat robots like contractors.
- Bake data rights into contracts. Specify ownership, derivative rights, and model-training permissions up front.
- Fund simulation. Make “no sim, no deploy” a rule.
The opportunity is to turn inconsistent, labour-heavy workflows into measured, repeatable services that improve every week. The competitive moat won’t be a single robot, it will be the end-to-end spine that runs from photons to policy – capture, label, simulate, deploy, verify, bill. AI is stepping into the physical economy. Buildings that prepare now will set the rules for everyone else inside their walls.
In Dr Marson’s monthly column, he’ll be chronicling his thoughts and opinions on the latest developments, trends, and challenges in the Smart Buildings industry and the wider world of construction. Whether you’re a seasoned pro or just starting out, you’re sure to find something of interest here.
Something to share? Contact the author: column@matthewmarson.com
About the author:Matthew Marson is an experienced leader, working at the intersection of technology, sustainability, and the built environment. He was recognised by the Royal Academy of Engineering as Young Engineer of the Year for his contributions to the global Smart Buildings industry. Having worked on some of the world’s leading smart buildings and cities projects, Matthew is a keynote speaker at international industry events related to emerging technology, net zero design and lessons from projects. He is author of The Smart Building Advantage and is published in a variety of journals, earning a doctorate in Smart Buildings.