Four agents. Built for the agentic stack.
Launching Q3 2026
ZenAI is the agent suite that runs inside ZenSpace. Each agent ships in recommendation mode and moves to controlled automation only where operators approve specific actions, set guardrails, and define limits. Launch-partner ready for Q3 2026.
4 agents
- Distribution
- Pricing
- Ops
- Alerts
Four agents. One mission.
Each agent is purpose-built for a single operational job. Together they replace the spreadsheets, the ops headcount, and the late-night Slack threads.
01
Distribution Agent
Multi-channel inventory synchronization
Reads
- PMS room status
- LiquidSpace / Deskpass demand
- Hotel website traffic
- Channel rate parity
Writes
Live inventory pushes to every channel
Rate parity corrections
Channel performance brief
Lost-business attribution
$ [14:02] +1 Pod-7 LiquidSpace · 14:30–15:30 · $48
$ [14:02] Synced inventory to 6 channels · 0 errors
$ [14:03] Parity drift detected: Mews $42 vs Site $45 → corrected
Watches
- Channel demand signals, parity drift, lost-business signals
Recommends / Acts
- Inventory push to channels, parity corrections
Business Outcome
- Higher capture rate from net-new channels
Guardrail
- Channel + rate-band whitelists set by operator; recommend-only on new channels until trust threshold met
↳ +22% capture rate from net-new channels
02
Dynamic Pricing Agent
Yield optimization per space, per slot
Reads
- Booking pace
- Lead-time distribution
- Segment mix
- Comp set pricing
Writes
Rate updates per space per slot
Demand elasticity curves
Override audit log
Forecast vs actual
$ [10:18] MR-4A 13:00–15:00 demand +38% · rate +12% ($45→$50)
$ [10:19] Cabana-2 weather penalty -8% · rate $80→$74
$ [10:20] Net daily lift est. +$612
Watches
- Booking velocity, occupancy curve, comp set
Recommends / Acts
- Hourly rate suggestions, soft-day discounts
Business Outcome
- Higher RevPOR on flex inventory
Guardrail
- Min/max price bounds set by operator; price moves require operator approval until trust threshold met
↳ +11% RevPAR on meeting spaces
03
Revenue Ops Agent
Daily brief, weekly forecast, monthly close
Reads
- Booking ledger
- Payment events
- Channel attribution
- Ops actions taken
Writes
Tuesday brief email
Weekly forecast PDF
Monthly close packet
Anomaly callouts
$ [08:00] Tuesday brief delivered to ops@hotel.com
$ [08:00] 4 actions queued · 2 high-priority
$ [08:01] Weekly forecast: +$8,420 vs same week LY
Watches
- Daily booking/revenue/utilization roll-up across spaces
Recommends / Acts
- Drafts morning brief for GM + asset owner; flags anomalies for review
Business Outcome
- One-page picture of yesterday + today before 8am
Guardrail
- Read-only — drafts and reports only. Never moves money or writes to systems of record.
↳ Less staff drag on month-end close
04
Alert Automation Agent
Anomaly detection across booking, sensors, payments
Reads
- Sensor telemetry
- Booking events
- Payment failures
- Channel error logs
Writes
Zendesk tickets on anomalies
On-call paging
Pre-cooling / HVAC commands
Refund / retry actions
$ [13:21] HVAC idle 47m · Pod-3 · ticket #4821 created
$ [13:22] Pre-cooling Pod-3 → 21°C by 13:50
$ [13:23] Payment retry success · Cabana-2 · $128
Watches
- Device health, access events, anomaly patterns
Recommends / Acts
- Routes alerts to the right staff member at the right time
Business Outcome
- Issues caught before guest experience is hit
Guardrail
- Configurable escalation matrix; alert thresholds operator-tunable; silent failures audited
↳ MTTR cut by 68%
Four agents. One decision graph.
Agents don't operate in silos. Every decision flows through the network — one agent's output is another agent's input.
- demand → rate ↓ ticket ↑ brief ← verify Distribution Pricing Ops Alerts
01
Distribution Agent detects demand spike on LiquidSpace.
+38% in 90-minute window
02
Pricing Agent recommends +12% rate uplift.
$45 → $50 · MR-4A · 13:00–15:00
03
Alert Agent verifies HVAC is in-range.
Pre-cool scheduled · 12:48
04
Ops Agent logs net-new revenue to morning brief.
Tuesday brief · +$612 lift
Operator-approved first. Controlled automation later.
ZenAI is built for properties that want the leverage without giving up the controls. Three rules govern every agent from day one.
Rule 01
Recommendation mode by default
Every agent ships in recommend-only mode. Operators see the suggestion. Operators choose to accept.
Rule 02
Permissioned actions
When automation turns on, it's scoped to specific actions, with limits set by the operator. Nothing happens outside the box you draw.
Rule 03
Audit log on every action
Every agent action is timestamped, attributable, and reversible. Nothing the agents do is opaque to the team running the property.
What ZenAI sees.
Agents are only as smart as the signals they read. ZenAI ingests from every operational surface, in real time.
- PMS feed ~600k events/day
- Channel demand ~120k searches/day
- Payment events ~30k events/day
- Smart-lock events ~80k events/day
- Occupancy sensors ~1.2M events/day
- Environment sensors ~400k events/day
- Website traffic ~20k sessions/day
- Calendar / weather hourly
Reserve your Q3 2026 launch slot.
Launch-partner properties get all four agents at the standard $99/room/mo Platform license — with direct line to the team building them. Limited capacity for the Q3 2026 launch.
Launch-partner benefits
- All 4 agents enabled at launch, no add-on fee
- Weekly product team office hours
- Co-development on custom agent behavior
- Public case study (optional, with approval)
Launch partners commit to a 90-day launch-partner revenue proof starting Q3 2026, with weekly check-ins. We prioritize properties with 6+ bookable spaces and an active PMS.