Red flags caught before reviews land.

Most negative reviews start with a single message a host never saw in time. Watchfox reads every inbound message — across every channel — for signs of trouble: an AC that isn't cooling, a missing key, a leak. When a red flag fires, the agent reroutes the thread and routes a work order before the review form opens.

What triage reads

Triage is a chat-aware layer that runs over the same inbox the messaging pipeline already reads from. It pulls four inputs every cycle:

• The full thread.Each message body, timestamp, and channel — so a guest who typed "the AC is on but it's not cooling" at 2 a.m. is caught and a guest who said the same thing in three short messages gets the same single read.
• The listing context. The unit, the previous clean, and any maintenance notes from the last turnover. A guest reporting a clogged drain two days after the previous clean is a different signal than the same report after a month.
• The red-flag ontology. A curated set of phrases and intents that Watchfox matches against the thread. Examples: AC not cooling, leaking faucet, missing keys, no hot water, broken latch, unreachable host, check-in failure.
• Your routing rules. Per listing, where to route a work order when a flag trips — SMS, Slack, email, or any combination. Default is SMS for after-hours and email for daytime.

How a flag is caught

When a new message lands, the agent runs four steps before the thread is ever opened in the inbox:

1. Tag intent. Tag the message as trouble / not-trouble / unsure using the ontology. This is what the marketing copy means by reads every message.
2. Score severity. A simple rule layer scores the trouble: low (cosmetic, can wait), medium (impairs comfort, needs prompt response), high (rebooking or refund territory, needs immediate work order).
3. Route the work order. Push to the channel and address your routing rules call for. SMS for medium and high; email for low; Slack if you have a channel mapped.
4. Edit the inbound reply. Promote the thread to urgent in the inbox and let the messaging pipeline draft a reply that acknowledges the trouble and sets expectations, instead of the routine response.

What shows up in the inbox

When a red flag trips, the thread in the inbox flips from awaiting you to urgent. The draft reply is rewritten to address the trouble — for AC that isn't cooling, the draft opens with the apology, the work-order reference, and a promise of triage within the hour, instead of the generic acknowledgment.

If the routing rules call for SMS, the SMS your vendors receive inside the same window is a single line — listing, severity, and a work-order id — so they can act without logging in.

Routing rules per listing

Most hosts route by time of day: SMS for after-hours, email for daytime. A few route per region: a different vendor gets the work order when the listing is in a specific zone. Both are editable on a per-listing basis from the dashboard.

What this looks like in week one

In week one, triage is reactive — Watchfox reads, scores, and routes. By week two you'll have enough data to tune the routing rules: which vendor responds fastest, which severity level gets SMS first, and which unit logs enough trouble to deserve a deeper property audit.

Related articles

Maintenance triage runs on top of the inbox: the same thread pipeline, the same draft reply, and the same audit log.