Agentic AI for
building operations.

Noda is agentic AI for commercial building operations: a virtual building engineer that works across an entire portfolio. Noda analyzes BMS trend data, meter data and weather around the clock to detect faults. Noda's AI agents then trace faults to their root causes and package the causes worth fixing as Projects, each carrying its estimated savings. Engineers review those Projects and decide which changes to make. In practice, an engineering team's hours go to changes with known savings, while Noda's AI agents carry the round-the-clock monitoring and diagnosis that few teams could staff across every building.

Fault detection and diagnostics (FDD) software stops at the fault list, while agentic AI carries on through diagnosis to a recommended change. A Lawrence Berkeley National Laboratory study of 317 buildings counted an average of 245 faults per building each month, and FDD leaves each of those faults for a person to investigate. Across a portfolio, that volume is far more than an engineering team can reasonably triage, so many faults go unresolved and the energy waste behind them keeps running. Noda groups related faults by their common root cause, estimates what fixing each cause would save, and delivers each cause as a Project ready for an engineer to accept. Handing engineers scoped work that is ready to execute means more savings get realized, faster.

Agentic AI reduces operating costs by finding operational waste that engineering teams rarely have the hours to diagnose: setpoints, schedules and control sequences that have drifted from what a building needs. Correcting that drift costs little or nothing, because each fix is a change to the controls. Left undiagnosed, the same waste shows up on every utility bill, and each month a fix waits is a month of savings a building never recovers. At one South Florida oceanfront resort, lowering the chilled water pump differential pressure setpoint and taking a third pump out of sequence cut 335,000 kWh a year.

Yes. Noda works with building management systems including Tridium Niagara, Johnson Controls Metasys, Siemens Desigo and Honeywell EBI. Each site gets a Noda LaunchPad, a secure on-premises gateway that connects to the BMS and to equipment controllers and IoT sensors running BACnet/IP, Modbus or MQTT. Meters connect through a separate Noda gateway. Deployment does not require replacing or reprogramming a BMS. Noda can therefore start finding savings in buildings as they run today, with the control systems and operating teams already in place.

Yes. Noda's GraphQL API makes equipment, points, trend history, Insights and Projects available to analytics and reporting software. Noda's MCP server connects AI assistants and agents that support the Model Context Protocol, giving each one scoped access to the same standardized building data that Noda's own AI agents use.

Noda acts on equipment automatically in one case: Automated Demand Management (ADM), which sheds or shifts site demand in response to a utility signal or an operational condition, on equipment configured and functionally tested for that purpose. Every other Project goes to a person to accept before anything changes in a building. Engineers stay in charge of how their buildings run, and ADM still captures demand savings during events that move faster than a person can respond.

No. Noda adds engineering capacity to the teams already in place. Noda's AI agents run the analysis that comes before a decision, including root cause analysis and savings estimates, and engineers decide which Projects to accept and carry out the work. As a result, the same team can cover more buildings, with more of each engineer's week spent completing fixes that carry known savings.