Permission-Native Dashboard Analytics — RBAC-Aware Metrics + AI Persona Attribution

[Weegy]

The Problem

Strapi v5.40.0 (released March 18, 2026) introduced a regression where users with multiple roles see duplicate entries in dashboard widgets — one copy per role assigned. Content counts are inflated, "Last edited" lists show duplicates. It's a trust-destroying bug that exposes a deeper architectural issue: dashboard metrics aren't computed with role-awareness at the query level.

The same pattern exists across competitors:

• Strapi — #25779: multi-role users see 3x data if they have 3 roles
• Directus — report queries can return duplicates when roles JOIN on collections
• Payload — no dashboard analytics at all (punt to external tools)
• WordPress — role-filtered views but no AI attribution or space isolation

No headless CMS today computes analytics that are natively aware of permissions, spaces, and AI agents.

---

The Proposal: Permission-Native Dashboard Analytics

Numen should ship analytics that are RBAC-aware from the SQL layer — not filtered post-query.

Core principles

1. Every metric respects role/space context — queries use role-scoped CTEs, never naive JOINs that duplicate rows
2. AI persona attribution — because Numen has first-class AI personas, analytics can show who (which persona or human) created, edited, published each piece of content
3. Space isolation by default — each space sees its own stats; admins get cross-space aggregate views
4. Pipeline metrics — not just "content count" but pipeline run success rate, AI generation time, review turnaround

Feature breakdown

Role-Scoped Content Metrics

• Content counts, draft/published ratios — all scoped to what the current user/role can access
• No duplicates when a user has multiple roles — DISTINCT on content ID, not role JOIN

AI Agent Attribution Dashboard

• "Persona A (GPT-4o) generated 47 articles this month, 38 published without edit"
• "Persona B (Claude) had 12 articles rejected in SEO review"
• Compare persona performance: generation speed, approval rate, edit distance from generated to published

Space Analytics (Enterprise)

• Per-space content velocity, pipeline throughput, publishing cadence
• Cross-space admin view with drill-down
• Space-level AI quota consumption

Content Lifecycle Metrics

• Average time from Brief → Published per content type
• Pipeline bottleneck identification (which step causes most rejections)
• Editor intervention rate (how often humans edit AI output before publish)

---

Why Numen Has Unfair Advantage Here

AI personas are first-class actors. Competitors have no concept of "which AI agent created this" — they just store content. Numen's pipeline tracks every step: which persona generated the brief, which LLM provider produced the content, which reviewer approved it. This attribution data exists already — we just need to surface it.

Space model enables clean scoping. Numen's multi-space architecture means analytics are naturally isolated. No query-level gymnastics needed to prevent data leakage between spaces.

---

Competitor Context

CMS	Analytics	AI Attribution	Role-Native Queries
Strapi	Basic dashboard (broken for multi-role)	None	❌
Payload	None built-in	None	❌
Directus	Insights module (basic)	None	Partial
Ghost	Basic (publisher-focused)	None	N/A
Numen	Full pipeline + AI attribution	✅ Persona-level	✅

---

Priority Signal

Ship before Payload or Directus adds analytics. Directus's Insights module is extensible — they could add AI attribution at any time. The window to be first with a CMS that can answer "which AI persona performs best for your brand" is open now.