Notification Preferences: Channels, Quiet Hours, and Digests

Configure how Activity & Alerts reaches you: which channels per severity, quiet hours, digest schedules, and per-category overrides. Critical and High always deliver immediately.

Activity & Alerts is one unified feed. Notification Preferences is where each user decides how that feed reaches them: which channels, which severities, which categories, and when the platform is allowed to interrupt them.

Preferences change delivery, not what gets recorded. Every event still lands in the feed and the Activity Log. Muting a category only stops the ping, not the record.

The delivery model

Every event has three attributes that decide whether you get notified:

1. Severity - Info, Low, Medium, High, Critical (see the unified severity ladder)
2. Category - Field Alerts, Weather, Scout, Tasks, VRA, Verification, Team
3. Field / group scope - which fields the event belongs to

Your preferences map each combination to a set of channels. The defaults are safe for most users; override anything that does not fit your workflow.

Critical and High severities always deliver immediately on at least one channel, regardless of your other settings. Quiet hours, digests, and category mutes cannot suppress them. This is a platform-level guarantee for field safety and yield protection.

Channels

Channel	Best for	Latency	Configurable per-user
In-app	Everything; the primary feed	Real-time	Always on
Email	Digests, summaries, verification receipts	Minutes	Yes
SMS	Critical field alerts when the app is closed	Seconds	Yes, phone number required
Push (mobile)	Field-time alerts, task assignments	Seconds	Yes, per device
Slack / Teams	Team-wide channels, ops rooms	Seconds	Yes, per workspace
Webhook	Downstream automation, farm management systems	Seconds	Yes, per endpoint

In-app is always on and cannot be disabled. It is your permanent record. All other channels are opt-in.

Default delivery matrix

These are the platform defaults. Every row is editable per user.

Severity	In-app	Push	SMS	Slack / Teams	Email
Critical	✓	✓	✓	✓ (channel + DM)	✓
High	✓	✓	✓	✓ (channel)	✓
Medium	✓	✓		✓ (channel)	Daily digest
Low	✓				Daily digest
Info	✓				Weekly digest

Setting up your preferences

1. Open settings

   Click your avatar → Notification Preferences. Preferences are per-user, not per-estate.

2. Verify your channels

   Confirm your email, phone number for SMS, and any connected Slack, Teams, or push devices. Unverified channels cannot receive Critical or High.

3. Adjust the severity matrix

   For each severity row, tick the channels you want. Critical and High require at least one channel; the UI enforces this.

4. Configure category overrides (optional)

   Some categories deserve tighter or looser routing than the severity default. Example: send every VRA lifecycle event to Slack for the ops room, but keep Team events in-app only.

5. Set quiet hours

   Choose a start and end time in your local zone. During quiet hours, Medium and below hold until the window closes. Critical and High always deliver.

6. Choose digest cadence

   Daily digest time and weekly digest day. Digests only include items below your immediate-delivery threshold.

Category overrides

Overrides let you route a single category differently from the severity default. Common patterns:

Ops room wants every VRA event

Category VRA → route all severities to #ops-room Slack channel. Individuals still get their normal severity routing on other channels.

Estate manager wants a quiet weekend

Category Team and Scout completions → digest only on Saturday and Sunday. Field Alerts and Weather stay on the normal schedule so real risks still ping.

Executives want headlines only

All categories → digest only, except Critical which routes to SMS. Result: one daily email plus SMS for anything urgent.

Agronomist on-call for a specific hazard

Rule card rice_blast → SMS on every severity for this user for the next 14 days. Everyone else keeps normal routing. Auto-expires.

Quiet hours

Quiet hours are a local-time window during which Medium, Low, and Info deliveries hold and release together when the window ends. Critical and High are exempt.

• Default: 22:00 to 06:00 local
• Weekend override: extend by up to 3 hours on Saturday and Sunday
• Travel-aware: quiet hours follow the device time zone, not your home zone. If you land in a different region, the window shifts automatically.

If your role includes an on-call rotation, the platform disables quiet hours during your on-call window automatically. This is set by the estate admin, not by the individual user.

Digest schedule

Digests are AI-summarized rollups of everything below your immediate-delivery threshold.

Digest	Default time	Contents
Daily	07:00 local	Last 24 hours of Low and Medium items, grouped by field, with the top actions surfaced
Weekly	Friday 16:00 local	Last 7 days of Info items, scout completion rate, VRA applications, verification acceptance rate

Digests link back to the feed for one-click drilldown. They never contain Critical or High items; those already delivered immediately.

Team and estate-level defaults

Estate admins can set recommended defaults for a team. New members inherit them on first login. Individuals can still override any row.

1. Open estate settings

   Estate → Notification Defaults.

2. Set the recommended matrix

   Adjust the severity × channel matrix that new members will start with.

3. Lock any rows (optional)

   For compliance or safety, admins can lock specific rows so individuals cannot disable them. Example: lock Critical → SMS on for every user. Locked rows show a padlock icon in the user’s settings.

What preferences do NOT control

• Whether an event happens - preferences never affect the risk engine, alert engine, or Activity Log.
• What Verification captures - every operation still cross-checks. See Verification Model.
• The unified severity ladder - severity is computed from the rule card, not from your preferences.
• Compliance floors - locked rows and the Critical / High delivery guarantee override any user setting.

Troubleshooting

I stopped getting SMS

Check that your phone number is verified. Carriers occasionally reject long-form SMS; the platform falls back to a shortened link. If the failure persists, the platform automatically escalates delivery to Push and Email and shows a red badge in your settings.

Slack channel is silent

The workspace token may have expired. Estate admin reconnects the workspace under Integrations. During disconnect, all Slack-only routes fall back to Email so you still get the message.

I am buried in Info-level notifications

Move Info to weekly digest only, and consider muting the Team category if you do not manage other users. Info to weekly is the platform’s recommended baseline; the default is Info-in-app for transparency, not for volume.

I want an audit of what was actually delivered

Settings → Delivery Log shows every send, including channel, timestamp, and delivery receipt (where the channel supports one). Useful for confirming that a Critical alert reached you before an incident.

Guardrails

This module follows the shared guardrails template. The agent and every non-agent write path must respect these rules.

Category	Rule
Input validation	Every preference row carries severity, channel, enabled, and (when scoped) field_ids[] or estate_id. Quiet hours require a valid IANA time zone; digest times require valid local times. Rule-card-scoped overrides require an existing rulecard_id and an explicit expires_at. Unverified channels cannot be set to receive Critical or High.
Preconditions	Channel must be verified before it can receive Critical or High. Estate-level defaults may only be set by an estate admin. Locked rows may only be modified by the admin who owns the lock or by an org admin. On-call windows may only be set by an estate admin.
Refusals	Disabling in_app on any severity. Disabling all channels for Critical or High (safety floor: they always deliver on at least one channel). Modifying an admin-locked row from a non-admin identity. Setting a rule-card override without expires_at. Muting a severity ≥ high signal beyond the bounded window declared by the safety floor.
Confirmations	Disabling a previously-verified channel for Critical or High (safety floor: refused unless another verified channel remains). Overriding estate defaults for a locked row (requires admin identity + reason). Setting quiet hours that span more than 12 hours. Enabling webhook delivery for Critical or High without an on-call fallback configured. Reversible changes (adding a channel, adjusting a non-locked row, changing digest time, editing a scoped override) commit optimistically with a 5-second undo.
Soft warnings	Preference change would silence more than 50% of the user’s current alert volume. Rule-card override expires inside the next 7 days. Quiet hours cover the estate’s on-call window. Preference conflicts with an estate default (informational; user can proceed).
Rate and scope limits	Bulk import of preference rows capped at N per org action (org-configurable) because of notification blast radius. No hard cap on individual edits; workflow volume is a soft warning, not a refusal. Webhook endpoint additions are rate-limited to prevent enumeration.
Audit	Every preference write conforms to the Audit Envelope. Module-specific fields: severity, channel, previous_enabled, new_enabled, scope (org | estate | user | field | rule_card), locked (boolean), expires_at (for scoped overrides), and override_reason (when overriding a locked row). Delivery routing changes are audited even when the underlying preference did not change (e.g., estate default propagation).
Escalation	Attempted disable of a safety-floor channel escalates to the estate admin. Repeated attempts to modify a locked row (three within 7 days by the same identity) escalate to the org admin. On-call window changes escalate to the estate admin for acknowledgement. Delivery failure on Critical or High escalates per Field Alerts.

Related

• Activity & Alerts - the unified feed these preferences route.
• Field Alerts - the four data-driven alert types.
• Activity Log - the timeline lens; unaffected by delivery preferences.
• Risk Model - defines severity, which drives routing.