Two clocks, per category
A useful SLA has two targets. Acknowledge: somebody has taken ownership. Resolve: it is done. A request that is acknowledged quickly and resolved late is a different problem from one nobody picked up.
Targets belong to categories, because a leaking pipe and a spare pillow are not the same urgency: housekeeping, maintenance, food, front desk. In a shop the categories change — a freezer that has stopped, an item out of stock — and the targets are minutes, not hours. What happens when a target is missed is the escalation.
An SLA only works if every request is timed from the moment it arrives. A request on WhatsApp or a word in the corridor has no start time, which is why a queue with a clock on it is the first thing an SLA needs.
Example, measured
At the pilot — one small property, 12–30 May 2026 — 45 guest requests were raised in nineteen days and 29 were resolved, with a median of 32 minutes from request to resolved. The other 16 were never marked resolved and are not counted. 28 were logged by staff, 14 came from a room's QR code, 2 by voice and 1 on WhatsApp.
That is also what honest measurement looks like: the requests that closed without a resolve time are left out and said to be left out, rather than counted as fast.
How Nivasik handles it
- Each request becomes a ticket with one owner and a clock: acknowledge and resolve targets per category, with room for a longer target at night, when the team is smaller.
- Requests arrive by QR sticker in the room — no app or login for the guest — or are logged at the desk.
- Targets are checked every minute, and a breached one escalates, re-assigned up the chain.
- Reports show volume, response and resolution time per staff member; the Android app rings a full-screen alert for critical requests.