AI Operations
An AI layer that runs the operation, not a chatbot bolted onto a calendar.
Propertory reads the live state of your portfolio, surfaces what is going wrong before a guest tells you, ranks it by consequence, and executes the routines your team has approved.
- Risk detection
- Daily prioritisation
- Governed Autopilot
- Full audit trail
No credit card required
Morning briefing
Portfolio · 64 units
- 07:12Turnover at riskDunaj 12 · 3B
Checkout 10:00 · arrival 14:00 · cleaner still assigned to two units
Reassign to standby cleanerAwaiting approval - 07:18Guest unverified 18h before arrivalPanorama · 21A
ID incomplete · access reveal blocked by policy
Send verification reminderAutopilot · guarded - 07:26City tax unpaidOld Town Loft
3 nights · 2 guests · settlement blocked
Issue payment requestAwaiting approval - 07:31Late checkout requestedRiverside 4
No same-day arrival · within add-on policy
Approved and charged as add-onResolved
Every proposal records its signals, its guard and its approver in the audit trail.
Inside the product
Autonomy you switch on deliberately, one routine at a time.
The most useful property of an AI operations layer is not speed. It is that a manager can always reconstruct what happened and why.
- 1Approval gatesEvery capability starts in propose-only mode. Execution is enabled per routine, per scope, by an administrator.
- 2Explainable decisionsEach proposal shows the signals used, the guard applied, the recommended action and the expected outcome.
- 3ReversibilityActions are recorded against the underlying records, so an incorrect step can be identified and corrected rather than silently absorbed.
- 4Health-aware holdsIf an operational data source degrades, dependent routines pause instead of acting on an outdated view of the portfolio.
How it works
One accountable trail, end to end
- 01 · Signals
Read the operational state
Reservations, verification status, payments, tax obligations, cleaning assignments, conversations and system health are all evaluated continuously.
- 02 · Detection
Identify what is drifting
Unassigned turnovers, arrivals without verification, unpaid balances, unanswered guests, blocked access reveals and data that has stopped updating.
- 03 · Prioritisation
Rank by consequence
Findings are ordered by how much damage they cause and how soon, producing a briefing your team can work top-down instead of triaging inboxes.
- 04 · Action
Execute or propose
Approved routines run inside their guards. Everything else becomes a proposal with a recommended action, ready for one-click approval.
Capabilities
What you get
Portfolio risk detection
Continuously flags reservations, turnovers, payments and guest threads that are heading towards failure.
Daily operations briefing
A ranked start-of-day view per role: what must be handled now, what can wait, and what has already been resolved automatically.
Context assembly
Pulls reservation, guest, property, payment and task history into one place so a decision does not require five lookups.
Message drafting and translation
Drafts replies grounded in the reservation record and the property's own information, in the guest's language.
Autopilot routines
Recurring operational sequences — reminders, verification chasing, task creation, status updates — executed within configured boundaries.
Safety guards
Limits on what may be sent, charged, changed or released, plus holds when underlying data is stale or incomplete.
Before and after
Disconnected tools versus one platform
Today, without Propertory
- Problems are discovered when a guest complains about them.
- Every morning starts by triaging inboxes to work out what matters.
- Automation is all-or-nothing, so nobody trusts it enough to switch it on.
With Propertory
- Drifting reservations, turnovers and payments are flagged before the guest notices.
- A ranked briefing per role: what must be handled now, and what can wait.
- Routines run inside approved guards, and everything else waits for a human decision.
Who it is for
What each team gets
Operations lead
One prioritised view of portfolio risk instead of five dashboards and a gut feeling.
Property manager
Proposals with the recommended action attached, ready to approve or reject.
Owner or director
An audit trail showing what was decided, by whom or by what, and when.
Questions
Frequently asked questions
Can the AI act without a human?
Only inside routines your team has explicitly enabled, within the safety guards you configure. Everything else stops at an approval gate and waits for a decision.
How do we know why the AI proposed something?
Every decision records the signals it used, the recommended action, the rule or guard that applied, who approved it and what the result was.
What happens if a data source is unhealthy?
System health monitoring surfaces stale or failing operational data, and affected routines can be held rather than acting on incomplete information.
Transparent pricing
€15 per property per month. Start with a 14-day trial.
AI is included under fair use. No credit card, no setup fee and no per-module upsell.