Chat has become the default interface for how we interact with technology. Sixty-seven percent of Americans now use a voice assistant at least once per month. Seventy-four percent of smartphone users engage with messaging apps daily. ChatGPT reached 100 million monthly active users faster than any consumer application in history. People are accustomed to asking questions and receiving answers through conversational interfaces. They expect to chat with their devices, their software, and increasingly their infrastructure. Yet the software that manages buildings and real estate portfolios remains dominated by dashboards, spreadsheets, and clicking through menus. The built world is lagging behind every other industry in adopting chat as the primary interface for accessing information and making decisions.

That gap is about to close. Nic Halverson, CEO of Occuspace, a building intelligence platform, is betting that chat will become the dominant way facility teams interact with their buildings. Occuspace launched Octi, a conversational AI tool designed to let facility managers, operators, and teams ask questions about their buildings and receive answers in natural language. The tool processes data from building systems like HVAC, lighting, and occupancy sensors. It integrates with existing building management software. When a user asks a question about energy usage or occupancy patterns or maintenance needs, Octi retrieves the relevant data and translates it into an answer. Halverson views this shift from dashboard-based interfaces to conversation-based interfaces as inevitable for building software.

The appeal of chat as an interface is straightforward. Dashboards require users to know what questions to ask before they encounter the data. They’re designed by engineers who make assumptions about what information matters and how it should be displayed. Different teams have different needs. Facility managers care about maintenance schedules. Energy managers care about consumption patterns. Building operators care about real-time system status. Space managers care about occupancy and utilization. Traditionally, each team gets their own dashboard built according to what engineers think they need. Halverson describes a different approach. “Everyone has a different personal relationship with their data and therefore their AI,” he said. “Now our software is not limited to what the engineers think we need.”

Chat interfaces allow each team to customize their own interaction with the data without requiring engineers to build separate dashboards. A facility manager can ask about maintenance trends. An energy manager can ask about consumption by zone. An operator can ask about system performance. Each query generates a response tailored to that person’s specific question. The interface adapts to how people actually work rather than forcing people to adapt to how engineers designed the interface. That flexibility is only possible through conversational interaction where users drive the inquiry rather than engineers predetermining what information gets displayed.

But chat interfaces also reveal a fundamental challenge. AI systems are excellent at gathering and processing raw data. They struggle with interpretation and context. Halverson articulated this tension directly. “AI is smarter than all of us but we have more domain specific knowledge,” he said. “Now we have to think about how we can get that knowledge into AI.” An AI system can identify that energy consumption spiked in a particular zone on a particular day. It cannot automatically know whether that spike indicates a problem, reflects normal variation, or signals an opportunity for optimization. That interpretation requires human expertise. The solution is using chat itself as a training mechanism. When people ask questions about their buildings, they reveal what they care about and what patterns they’re trying to understand. That dialogue teaches AI systems what matters and how to contextualize raw data within domain-specific knowledge.

Over time, the relationship between human expertise and AI capability evolves. Initially, humans ask questions and AI retrieves data. Then humans and AI collaborate to interpret that data. Eventually, humans are mostly there to make decisions that require judgment. The AI handles information gathering and pattern recognition. Humans provide context and make calls on actions that involve organizational priorities or risk tolerance. Halverson described what that looks like at the decision layer. “Right now we want to be able to ask it ‘what should I do?’ Eventually that will go away and it will just automatically tell you what you likely need to know.” The progression is from humans asking questions, to humans reviewing recommendations, to humans making decisions based on AI-provided intelligence that has learned to anticipate what they need to know.

That progression requires a context layer. AI systems need to understand not just the data but the environment in which decisions get made. Building occupancy matters differently at different times. Energy consumption matters differently based on weather. Maintenance needs matter differently based on asset age and criticality. AI trained only on raw data without that context will miss the nuance that makes recommendations actually useful. Halverson’s view is that the context layer comes from dialogue. When facility teams chat with their building systems and ask questions, they’re implicitly teaching those systems what context matters. Over time, AI learns not just to answer questions but to anticipate the context in which questions will be asked.

The deeper shift is about what people actually want from their software. Most facility managers and operators don’t want to be bothered with data. They want answers to their problems. If energy consumption is optimal, they don’t need a report about energy. If equipment is functioning normally, they don’t need alerts about performance. If occupancy is within expected parameters, they don’t need occupancy dashboards. The only time people want to see data is when something requires their attention or decision. Chat interfaces facilitate that shift. Instead of pulling up dashboards and scanning for problems, people ask their building systems what they need to know. The system, trained on what matters and what context applies, delivers answers rather than raw data. The goal is for facility teams to spend less time looking at screens and more time making decisions that require human judgment.

Building software will eventually evolve the same way consumer software has. People stopped using websites when mobile apps became convenient. They stopped using apps when voice assistants became reliable. Each transition reduced the friction between wanting information and getting it. Chat is the latest reduction in that friction. You don’t need to open an application. You don’t need to navigate menus or find the right dashboard. You ask a question. The answer comes back in natural language. If that model works for consumer software and enterprise software, it will work for building software. The companies building the next generation of facility management platforms will be the ones that prioritize conversation over dashboards.

The post The Next Evolution of Building Software Interfaces Is Conversational appeared first on Propmodo.

​