Story 012: UC-08 CartLoop Commerce: Personalization and Destination Trust (governance compliance 4/4)

Campaign Trigger & Audience

Activate personalization and destination trust when CartLoop Commerce schedules an audience-specific campaign burst.

Marketer Actions (UI + automation)

• Create an audience segment with clear campaign goal tags.
• Configure destination rules and preview each segment branch.
• Enable automated QA checks before go-live.
• Schedule launch window and monitor live KPI dashboard.

Expected Outcome

CartLoop Commerce delivers trusted personalized destinations that improve campaign conversion quality.

KPI Watchlist

• audit_completeness_pct
• approval_sla_minutes
• policy_violation_escape_rate

Tracking Setup (events, attribution tags)

• Track uc-08.configure-personalization.policy_check_started
• Track uc-08.configure-personalization.audit_event_recorded
• Track uc-08.configure-personalization.policy_check_passed
• Attach attribution tags: campaign_id, audience_id, destination_variant.

Failure Signals + Recovery

• Signal: manual domain decision overrides automated risk score.
• Recovery: pause affected segment and re-run destination preview.

Rollback/Support Handoff

• Rollback: execute revert_rule.
• Support handoff: include story ID, segment ID, and event timeline.

Acceptance Checklist

• Audience segmentation reviewed.
• Destination mapping validated.
• Tracking and attribution tags verified.
• Rollback owner confirmed.

Gitagent Grounding Context

• Description: Autonomous external ChatGPT/OpenAI OAuth MCP agent for handle-scoped 10x.in workflows, connector recovery, and delegated control-plane execution
• Autonomous external ChatGPT/OpenAI OAuth MCP agent for handle-scoped 10x.in workflows, connector recovery, and delegated control-plane execution
• I am the development agent for 10x.in — a production-grade, multi-tenant link shortening and creator economy platform deployed on AWS. I understand the full stack from CDK infrastructure to CloudFront edge routing, from DynamoDB data modeling to React frontends, and from MCP tool orchestration to Temporal workflows.
• - API: Monolithic Lambda handler with 277 method-routes across handle-scoped, public, admin, and internal paths
• Key domain entities: Handles, Links, Visitors, Campaigns, Pages, Profiles, PersonalizationRules, ChainRules, TrackingTemplates, Skills, AgentRuns.
• - Deployment reliability: Follow the established deployment workflow; use IAM wrappers
• - Follow the skill-based workflow: match task to skill, read the skill file, follow the checklist
• - Use scoped commit messages: feat:, fix:, docs:, chore:
• | /v2/handles/{handle}/* | JWT | Handle-scoped operations |
• | Commerce & Growth | 8 | PersonalizationRules, Campaigns, Pages, Visitors, BillingEvents |
• ## Deployment Workflow
• ## Known Failure Patterns
• Map external end-user prompts to the canonical workflow, tool family, scope set, and rollback path.
• Use this skill when an external ChatGPT/OpenAI prompt needs to be converted into one of the canonical handle workflows.
• workflow: configure-personalization
• required_scopes: mcp.connect, tracking.personalization.read, tracking.personalization.write, routing.context_origins.read
• rollback: revert_rule
• failure_paths: preview-inconclusive, validation-regression