Create a Multi-Store Operations Command Center in Enter for a U.S.-based restaurant brand with 24 stores. Help headquarters publish SOPs, training, and campaign instructions as store-level tasks, verify execution with evidence, and route store issues to an owner before they stall.

[TARGET USERS]
- Director of Operations (HQ): Publish updates, monitor completion and exceptions, and decide escalated issues.
- Regional Manager: Verify or return store submissions, own regional issues, and follow up on late stores.
- Store Manager: Acknowledge updates, complete steps, upload evidence, raise issues, and request extensions.
- Administrator: Manage stores, regions, timezones, users, data imports, and access.

[CORE WORKFLOW]
1. Publish an SOP update, training module, campaign instruction, or performance follow-up to selected stores. Create one task per store with acknowledgement, steps, required evidence, and a due time in the store's local timezone.
2. Let stores acknowledge, complete steps, upload evidence, and submit. Managers verify or return each submission with a reason.
3. Let stores raise issues linked to a task with category, affected area, and requested support. Triage severity, assign an owner, and run an SLA clock that pauses while waiting on the store.
4. Escalate issues to the HQ Action Queue when no regional owner acts within the SLA. Mark linked tasks as blocked and extend their due time for the blocked period.
5. Publishing a new version archives the previous one and cancels its unfinished tasks while keeping verified history; stores only see the current version.

[CORE PAGES]
- Operations Dashboard: Store completion, overdue tasks, awaiting acknowledgement, open issues, critical and SLA-breached issues, live HQ updates, and late stores.
- HQ Action Queue and Issue Board: Escalated, unassigned, SLA-at-risk, and verification-backlog items with next actions.
- Publishing Center: Updates, versions, store completion matrix, exceptions, and verification actions.
- Training Library and Performance Follow-up: Modules, assignments, quiz results, gaps by store, and open follow-ups.
- Data Management: Imports, connections, stores, users, and demo data.

[DATA AND BUSINESS RULES]
- Use stable IDs to link updates, versions, store tasks, steps, evidence, issues, SLA events, training assignments, and users.
- Completion = verified store tasks / non-cancelled store tasks. Submitted tasks do not count until verified.
- Judge overdue status in each store's own timezone; blocked time extends the due time, and waiting on the store pauses the SLA.
- AI assistance may summarize issues and draft follow-ups, but must not verify tasks, close issues, or change severity on its own.
- Include a demo case: "Allergen labelling SOP v4" adds sesame as a named allergen and requires a new laminated matrix at every prep station. Published to 24 stores, it has 5 of 24 tasks verified (20.8%), 18 overdue, 4 not acknowledged, and 3 blocked by an issue. Oakland Broadway (1005) was returned with 2 of 4 steps done; its issue "the v4 matrix lists sesame under seeds but the label printer still prints the old category" is Critical and escalated to HQ after missing the SLA. Philadelphia Rittenhouse (1023) submitted all 4 steps and awaits verification.

[VISUAL DIRECTION AND DELIVERABLES]
- Use a dense but readable command-center layout with filters, status chips, SLA timers, and drill-down from every number.
- Support desktop and mobile so store managers can acknowledge, upload evidence, and raise issues from the floor.
- Use labelled demo data with consistent IDs, store numbers, timezones, and U.S. date conventions.
- Deliver a publishable demo with one traceable path from an HQ update to store evidence, an escalated issue, and manager verification.

Before building, provide a requirements summary, missing information, page architecture, data model, completion and SLA rules, role permissions, state transitions, and acceptance criteria. Do not begin generating pages yet.