# Notifications Notifications are how the platform reaches you when something needs attention — fast, in the right channel, with enough context to act. They are configurable per user and per **organization** (an isolated workspace, typically one agency / chain / brand whose data does not mix with anyone else's), scoped to what you actually want to be told about. ## What kinds of notification exist | Kind | Examples | Default channel | | -------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------------------------------- | | **Operational alerts** | Failed channel writes, partner connection errors, mapping gaps | In-console toast + email | | **Financial events** | Chargeback opened (e.g. EUR 320 reversed on booking BK-12345), settlement failed, refund pending action | Email + in-console | | **Contract events** | Contract expiring in 30 days, signature pending more than 7 days, contract amendment ready for review | Email | | **Adragent events** | Staged operation needs your approval, staged operation expired (e.g. *"+10% BAR on Hotel Sunrise"* not approved within 15 min), Adragent run failed | In-console + WhatsApp/SMS for time-sensitive | | **Catalog events** | Property published, content review queued, image upload completed | In-console only | | **Operational alerts (high severity)** | Sold-out for the next 7 days on a high-revenue property, booking pace 3× the 4-week average, mass-cancellation event (e.g. 12+ cancellations in 1 hour) | All channels (in-console + email + SMS) | You can choose which events reach which channel. ## Channels you receive on * **In-console toasts and inbox** — your daily channel; appears as you work. * **Email** — for things that should outlive the session (action items, summaries, approvals). * **WhatsApp** — for off-desk approvals and time-sensitive checks (paired with Adragent's WhatsApp surface). * **SMS** — for the most time-sensitive — failed channel writes during a peak period, chargebacks, contract expiring tomorrow. Each channel is opt-in per kind. You decide what reaches you where. ## How notifications group A single event often produces several notifications across different audiences: * A chargeback opens → finance gets an email + an in-console alert; the booking owner sees a toast next time they open the booking. * A channel goes unhealthy → ops sees an in-console alert; if it stays unhealthy for more than 4 hours, the organization admin gets an escalation. * A contract expires next week → the responsible teams see it; the property owner sees a toast on the property page. ## Severity and quiet hours Each notification has a severity level — info, warning, critical. Severity decides: * Whether the notification ignores quiet hours. * Which channels it reaches (critical may use SMS even if the user has SMS off for normal events). * Whether the platform escalates after a window of inaction. You configure quiet hours per user (no SMS between 22:00 and 08:00, no WhatsApp on weekends). Critical notifications can override quiet hours; everything else respects them. ## Acknowledgement and follow-up For action-required notifications: * The notification has an *acknowledge* action — you confirm you saw it and are working on it. * Unacknowledged notifications surface again after a configurable interval. * Repeated non-action escalates to a higher role (a chargeback unanswered for 5 days alerts the organization admin). This is what stops important things from falling through the cracks during high-volume periods. ## Adragent and notifications Adragent has access to notifications as both a producer and a consumer: * **Producer** — when Adragent stages an action that needs approval or detects an anomaly worth flagging, it can produce a notification on the operator's chosen channel. * **Consumer** — common phrases that map to operations on notifications: * *"Show me notifications I haven't acted on in the last 24 hours."* * *"Find every chargeback notification still open after 5 days and escalate."* * *"Mute settlement-completion notifications for the next two weeks — I'll check the dashboard instead."* Read operations return immediately; configuration changes preview before applying. ## Configuration You configure notifications at three levels: * **Per user** — your channel preferences, quiet hours, severity overrides. * **Per organization** — overrides for an entire organization (e.g. *all operators in this organization get SMS for chargebacks*). User settings override organization settings. ## Where to next * **Search past platform events** → [Activity log](/console/notifications-and-ops/audit-log.md) * **The Adragent surface that produces and consumes notifications** → [Adragent overview](/console/adragent/overview.md) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://adrasis.gitbook.io/console/notifications-and-ops/notifications.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.