Prompt 13 — Survey campaign, rater, notification, and release lifecycle

Required context

• Program builder and question assembly engine
• Rater relationship and anonymity policies
• HR and manager workspaces
• Current survey lifecycle screen

Prompt

Implement a human-readable campaign lifecycle that replaces raw state and rater JSON with safe operational workflows.

1. Campaign model and state machine

Document and implement approved states such as:

draft → nomination → review → scheduled → inviting → collecting → closing
→ processing → threshold_wait → ready_for_release → released → archived

Include paused, cancelled, failed, reopened, and expired paths where valid. Do not expose internal job states as primary user language.

Every transition defines:

• Allowed actor/permission
• Preconditions
• Side effects/jobs
• Idempotency
• Notification behavior
• Audit event
• Reversal/compensating action
• User-visible status

2. Campaign overview

Show:

• Program and version
• Population/scope
• Observation and collection periods
• State stepper
• Participants and exclusions
• Rater-group readiness
• Completion funnel
• Threshold readiness
• Delivery/job errors
• Next recommended action
• Activity history

3. Participant and rater management

Provide named, filterable tables for authorized users:

• Subject
• Manager
• Proposed/approved raters by relationship
• Internal/external status
• Invitation status
• Completion state shown only within anonymity policy
• Conflicts, duplicates, missing groups
• Exceptions and reasons

Use people names and relationship labels, not user IDs or JSON.

4. Nomination and approval

Support tenant-configured patterns:

• System proposal
• Subject nomination
• Manager approval
• HR exception
• External raters
• Replacement before/after invitation according to policy

Preview group size and anonymity impact. Audit overrides.

5. Notifications and reminders

Implement templates and jobs for:

• Invitation
• Reminder
• Deadline change
• Completion confirmation
• Report ready/released
• Recovery/error where appropriate

Provide schedule, channel, localization, send status, failure, retry, suppression, and tenant branding. Ensure idempotency and no duplicate reminders.

6. Closing, processing, and release

• Validate threshold and completeness.
• Explain pending/suppressed groups.
• Run report/import/AI processing through observable jobs.
• Allow authorized preview of aggregate output without breaking privacy.
• Release to subject according to policy.
• Notify and route to reflection/PDP.
• Support archival and retention.

7. External-rater experience

Use a secure, mobile-first token/auth flow with privacy notice, estimated time, save/resume policy, and completion. Guest raters do not require paid subject seats.

8. Tests

• Every state transition and invalid transition
• Concurrent actions/idempotency
• Invitation token expiry/replay
• Notification retry and duplicate prevention
• Rater replacement
• Threshold and suppression
• Scope/role access
• Cancellation/reopen
• Timezone/deadline behavior

Acceptance gates

• No raw JSON is shown in normal campaign operations.
• HR can identify the next action and failure reason.
• Relationship completion and thresholds do not enable re-identification.
• Notification and transition side effects are idempotent and audited.