In-house housekeeping app

The plan is to own the housekeeping layer end to end, retiring the paid app and running it PMS-native, property by property.

The portfolio pays for its housekeeping app as an annual subscription. More than the cost, I want to own the layer (the scheduling, the room-state write-back, the data) instead of renting it, especially now that the AI co-pilot around it is already mine.

How it works

The target is a PMS-native system: a housekeeper mobile app plus room-state write-back through the Mews Connector API. Read room state and reservations from the API, write clean, dirty, and inspected back, and use the real-time space-event WebSocket (webhooks lag a few minutes) with reconnect and gap-fill mandatory.

Sequencing: agent layer first (the co-pilot, already live), then the mobile app and write-back, then property-by-property cutover
Every cutover reversible, with a shadow-run and burn-in against a tested manual fallback before disabling the paid app anywhere

The honest read is that the cost saving is not the case. Real payback lands past three years once you load the build and perpetual on-call. The real case is the differentiators the paid app lacks (occupancy-driven scheduling and Spanish-native AI chat) plus owning a Mews-native layer.

Planned and gated, not committed. It only proceeds past a live write-back spike on a real token and a true NPV. That discipline is the point. The company of one is not build-everything. It is build the seat where owning it actually compounds, and be willing to shelve the rest.