HVAC AI Agents home

Private beta · Hospitality

Occupancy-aware HVAC for every guest room and back-of-house.

AI agents that pull from your PMS, watch every room + BOH unit, and drop HVAC spend the moment occupancy drops. Without compromising on comfort scores.

Join the waitlist →

15-25%

designed room HVAC reduction

when occupancy drops, not before

comfort

score uplift

guests can't tell the agent is there

on-edge

no guest data leaves property

PMS bridge stays local

Three modes — autonomous leads here

Primary for this page

Autonomous, occupancy-aware

Reads occupancy from your PMS, adjusts room HVAC the moment a guest checks out, recovers before the next check-in.

Engineering copilot

Engineering asks: 'why is wing 3 hot?' Copilot answers + opens the ticket.

MCP for your stack

Connect your in-house LLM or third-party agents via MCP.

Hotel and restaurant HVAC

Hospitality HVAC AI — PMS-aware setbacks that protect brand comfort standards.

HVAC AI Agents connect to your Property Management System — Oracle Opera, Mews, Cloudbeds — so room temperatures adjust on guest checkout and pre-cool before the next reservation, cutting kWh/occupied-room-night without triggering comfort complaints or violating brand SLAs.

The hardest tradeoff in hospitality HVAC is occupancy-based setbacks: set them too aggressively and the next guest walks into a 78°F room; set them too conservatively and you're paying to cool empty rooms around the clock. Most hotels resolve this by leaving setbacks shallow or disabling them entirely, which means HVAC runs at full comfort settings whether the room is occupied or not. That's the right call for guest satisfaction, but it leaves 20–30% of room HVAC energy on the table.

The agent solves the tradeoff by reading checkout and reservation data from your PMS in real time. When a guest checks out, the room enters a deep setback immediately. When the next reservation is 90 minutes out, the agent pre-conditions the room so it hits setpoint before the new guest arrives. Guest comfort SLA — room temperature within ±1°F of setpoint at check-in — is monitored continuously and reported alongside the energy savings, so the facilities team can show that setbacks did not cost comfort points.

Back-of-house is a separate problem. Hotel kitchens, laundry rooms, and mechanical spaces often run HVAC continuously because there's no schedule integration — the exhaust fan runs overnight even when the kitchen closed at 11pm. The agent applies BOH-specific schedules based on POS close times, linen department hours, or configurable operating windows, trimming runtime in spaces where occupancy is predictable. For resorts and seasonal hotels, an off-season mode applies deep setbacks chain-wide with automated weekly health checks to confirm no equipment has drifted into a fault state.

Where it pays off

Concrete scenarios from hotel, resort, and restaurant operations.

Four patterns we see repeatedly across hotel chains, resort properties, restaurant groups, and hospitality REITs.

Hotel chain ops director

Brand comfort standards require room temps within ±2°F of setpoint at check-in. Aggressive setbacks save energy but generate front desk complaints. Conservative setbacks cost $800K/year in avoidable HVAC spend across the portfolio.

→ PMS-integrated setbacks pre-condition rooms 90 minutes before check-in. Guest comfort SLA monitored continuously. Energy savings and SLA pass rate reported together — proving setbacks didn't cost comfort scores.

kWh/occupied-room-night vs. SLA pass rate

Resort facilities manager

Off-season occupancy drops to 15%. HVAC runs at full comfort settings because there's no automated off-season mode. Energy bills stay near peak-season levels even when 85% of rooms are empty for weeks.

→ Off-season mode activates chain-wide based on occupancy thresholds. Deep setbacks apply to vacant wings. Weekly automated health checks confirm no equipment has drifted into fault. Energy costs track occupancy instead of the calendar.

Energy cost proportional to actual occupancy

Restaurant group ops director

Kitchen HVAC — exhaust, makeup air, and cooling — runs 24/7 at most locations because no one integrated it with the POS close schedule. Overnight runtime in a closed kitchen is pure waste.

→ Kitchen HVAC follows POS close and open times, trimming overnight runtime by 5–7 hours per location. Exhaust and makeup air step down 30 minutes after last ticket; step up 45 minutes before first prep shift. Fuel and electric savings visible per location.

Kitchen HVAC runtime tied to actual operating hours

Hospitality REIT asset manager

Portfolio energy benchmarking across 40 properties in three brands requires quarterly manual data pulls from each property's BMS and utility bills. Outlier properties aren't visible until the quarter closes.

→ kWh/occupied-room-night rolls up across the portfolio in real time. Outlier properties surface within days of onset. Asset manager can see which properties are above brand-cohort benchmarks without waiting for the quarterly report.

Portfolio energy benchmark updated in real time

FAQ

Hospitality HVAC AI — common questions.

  • How does the PMS integration work with Oracle Opera, Mews, or Cloudbeds?

    The agent connects to your PMS via its standard API — Oracle Opera REST, Mews API, or Cloudbeds API — and subscribes to checkout and reservation events in real time. On checkout, the room enters a configurable deep setback immediately. When the next reservation is within a configurable pre-conditioning window (default 90 minutes), the agent starts conditioning the room to setpoint. No changes to the PMS are required. Read-only by default; write-back to thermostats is opt-in per property. Contact us with your PMS platform and version to confirm the integration path.

  • How do setbacks balance against brand comfort standards like Marriott or Hilton SLAs?

    The agent treats your brand comfort SLA as a constraint, not a tradeoff. You define the minimum pre-conditioning lead time and the arrival setpoint target; the agent works backward from the reservation to schedule pre-cooling. Guest comfort SLA — room temperature within ±1°F of setpoint at check-in — is monitored continuously and reported as a pass rate alongside energy savings. If a room fails to reach setpoint in time, the alert fires before check-in so front desk can respond. Properties typically maintain 97–99% SLA pass rates after the first two weeks of tuning.

  • What is off-season mode and how does it work?

    Off-season mode applies deep setbacks to rooms and common areas when property occupancy drops below a configurable threshold — typically 20–25%. Each vacant room holds a wider temperature band (e.g., 60–85°F) instead of a comfort setpoint, cutting conditioning energy by 50–70% per room. Weekly automated health checks verify no equipment has drifted into fault, no pipe temperatures are approaching freeze risk, and humidity is within bounds. When occupancy rises back above threshold, the property returns to normal operating mode automatically. Contact us about seasonal configuration for your property type.

  • How does the agent handle kitchen, laundry, and back-of-house HVAC?

    BOH spaces get schedules independent of guest room logic. Kitchen HVAC — exhaust hoods, makeup air units, and supplemental cooling — follows POS close and open events or a configurable operating window. Laundry follows the linen department schedule. Mechanical spaces use a minimal-runtime ventilation schedule outside occupied hours. Each BOH zone has its own setback floor so equipment health isn't compromised. Runtime savings in kitchens typically run 5–7 hours per day per location. Ask us about BOH schedule setup during the pilot.

  • Does it work with PTACs and terminal units — Amana, Friedrich, Sanyo?

    Yes. The agent has production integrations for Amana, Friedrich, and Sanyo PTACs alongside central-system brands like Carrier, Trane, Daikin, and Mitsubishi. For PTAC-heavy properties, the agent reads each unit's reported temperature and runtime via the control network or the brand cloud API, and write-back setbacks are applied per unit. Mixed properties — PTACs in guest rooms, central AHUs in common areas — are supported in the same deployment. Check /coverage for model-level support status, or contact us with your specific unit models.

  • What does a multi-property rollout look like?

    Three phases. Phase 1: one to three properties as a pilot, focusing on PMS integration validation and setback tuning over 30 days. Phase 2: regional expansion across 10–20 properties with the ops team owning daily monitoring. Phase 3: portfolio-wide rollout with automated energy benchmarking and off-season mode enabled. Properties are onboarded in batches of 5–10 with a standard integration checklist. Most pilots are operational within two weeks of kickoff. Request a pilot scope from the form below.

  • How do you bill for hospitality deployments?

    Per property per month, with volume tiers starting at five properties. No per-room, per-unit, or per-point metering. The base subscription includes PMS integration, room setback scheduling, guest comfort SLA monitoring, and BOH scheduling. Off-season mode and portfolio benchmarking views are included in multi-property plans. A single-property pilot is a fixed-fee engagement. Annual contracts include onboarding, PMS integration support, and setback tuning. Request pricing from the form below — quote depends on property count and PMS platform.

Speaks to your existing kit

Carrier, Trane, Daikin, Mitsubishi, LG, Lennox, York, Samsung — 20+ HVAC, home-automation, and BMS brands.

63 brands across 3 categories — HVAC (31), Home Automation (18), BMS (14). Protocols: BACnet, KNX, MQTT, Matter, Modbus, REST, WebSocket, Z-Wave, Zigbee.

See full coverage →

How it stays out of your way

Secure

Sealed data plane. Per-site auth. Audit log on every setpoint touch.

Runs on the edge

Deploys at the building edge — your data doesn't leave the site to be useful.

BYO LLM

Works with Claude, ChatGPT, and any MCP-compatible client. You pick the brain.

Private beta

Pilot it on one property first.

Designed for hotels, restaurants, casinos, and resorts.