For more than two decades, the smart home industry has focused on a single problem: how to make devices easier to control. From voice interfaces to mobile apps and automation platforms, each layer improved how users interact with connected devices.
Yet these advances have largely focused on improving inputs, rather than reducing the need for user involvement. Smart homes still depend on users to define what should happen, when it should happen, and how different systems should work together.
As large language models reshape how people interact with software, expectations are changing. People are becoming less interested in issuing step-by-step instructions and more interested in expressing intent and having systems carry that intent through to completion.
That shift matters everywhere, but it may matter most at home. For all the energy poured into AI over the past few years, most products have been built for two groups: technical users or professionals trying to move faster at work. But life at home doesn’t work like that. It’s not a clean workflow. It’s a constant layer of small, overlapping decisions — schedules shifting, groceries running low, kids needing attention, plans changing at the last minute. And in most households, all of that still has to be held together by someone.
The smart home is already mainstream. More than 40% of households globally now own at least one smart device, with adoption in markets like the U.S. exceeding 80%.
The market continues to scale rapidly, projected to grow from over $120 billion today to more than $500 billion by the end of the decade. But scale has not solved the core problem. As devices and platforms expand, the experience remains fragmented, spread across different apps, ecosystems, and interfaces.
At the same time, AI itself is evolving. The global AI agent market is expected to grow at more than 40% annually, reflecting a shift from systems that respond to commands to systems that can act independently. That shift is beginning to change how AI operates in the home. Most systems today still follow a familiar loop: request, response, confirmation. Even when connected to external services, they remain step-based and user-driven.
What is emerging instead is a different model. Rather than assisting decisions, systems are starting to execute outcomes. One early example of that shift is SuperNori, developed by Domus Next, a San Francisco company founded in 2025. The company’s broader premise is that the home should not be treated as a collection of smart features, but as a system-level coordination problem. Public materials around the product consistently position it less as a family app and more as shared AI infrastructure for household life.
That framing is important. The breakthrough is not that AI can now talk to your home. It is that AI may finally be able to operate across software, services, and hardware in a way that feels coherent to the people living inside it.
If execution is the new expectation, the structure of the system itself has to change. Most smart home products today are still built as interfaces, sitting on top of devices and exposing controls through apps or voice commands. That model works for issuing instructions, but it breaks down as environments become more complex.
What SuperNori represents is a different architectural approach. It is designed as a space-native home operating system, built through deep integration with Android system-level permissions and the Home Assistant ecosystem. Rather than acting as another interface, it operates across software, services, and hardware as a unified execution layer.
That shift changes what “control” means in the home. Instead of telling devices what to do step by step, users express their intent, and the system determines how to execute it across environments. Devices, applications, and third-party services are coordinated within a single workflow rather than as separate endpoints.
At the core of that system is a different way of handling tasks. SuperNori translates natural language into structured actions, breaking requests into multiple steps and executing them in parallel across digital environments and physical devices. It continues running in the background until the task is fully completed, rather than stopping at suggestions or confirmations.
At the software layer, SuperNori can directly operate user interfaces, navigating screens, typing, and completing tasks the way a human would without relying on APIs or prebuilt integrations. At the hardware layer, it extends that same capability into the physical world through Home Assistant, coordinating devices across protocols and brands as part of a single system.
The difference shows up most clearly in everyday life. In traditional systems, tasks are fragmented. Users move between apps, trigger routines, and manually connect steps across digital and physical environments.
In this new model, those steps are absorbed into the system itself. A system might notice unusual traffic before anyone is awake and suggest booking a ride early. It sees pantry staples running low and lines up options before anyone opens a shopping app, factoring in what the family usually buys and what’s already on hand. It picks up on a child’s upcoming science topic and pulls together something age-appropriate in advance, like a few simple explanations or materials they can actually use. And when plans start to fall apart, like a last-minute conflict before an anniversary dinner, it surfaces workable alternatives before the situation turns stressful, based on timing, location, and past preferences.
What changes is not just convenience, but continuity. Instead of reacting to individual requests, the system carries context forward — adjusting schedules when conflicts arise, keeping plans aligned across people, and handling routine coordination in the background.If AI is going to matter outside the office, this is one of the clearest places it has to prove itself.
If the last decade of AI was defined by models, the next may be defined by systems that can act.
“We don’t think the future of AI lives in the chat box,” Isaac Long, Co-founder of Nori said in a recent internal discussion. “Models are becoming infrastructure. What matters now is whether a system can actually follow through, perceive what’s happening, make decisions, and carry them out in the real world.”
That shift is already pushing differentiation up the stack toward orchestration, context, and execution.
For years, AI has been optimized for work. The next phase will test something harder: whether it can handle everyday life, where coordination is continuous and rarely structured.
Systems like SuperNori point to one possible direction. Not as a finished answer, but as an early signal of a broader shift from tools that respond to requests to systems that take responsibility for outcomes.If that holds, the real impact of AI may be measured not by how well it answers questions, but by how much of the invisible work of running a household it can quietly take off people’s hands.
Learn more about SuperNori and follow the development of Family AI here.