AI Content Command Center
A controlled content operating workflow for Qatar Living's Events, News & Social team — not a generic AI writer. The system detects the opportunity, prepares the content, reviews governance, routes it for human approval, then measures and learns. AI prepares the work; humans approve every decision.
What this system proves
- It writes in the Qatar Living voice. The brand model is derived from a study of 1,999 published articles: entity-first headlines, attribution-heavy prose, the institutional stance verbs the newsroom actually uses.
- It enforces governance automatically. 46.5% of the analysed content needs senior review; the system routes those items to humans by default and never auto-publishes.
- It tidies copy at publish time. The same boilerplate CTA paragraph was detected in ~1,425 articles, alongside paste artefacts and stray markup that hurt SEO, accessibility and machine-readability.
- It surfaces the event-freshness problem on day one. 57.6% of listed events had already ended at study time (80.5% today); the system flags every one with a suggested correction for approval.
Live News Radar
The pilot continuously scrapes real Qatar news (QNA, The Peninsula, Qatar Tribune, Gulf Times and more) and classifies each item for governance risk. Pick any story and the system re-reports it in Qatar Living's tone — reworded, governed, queued for approval. Only headline + source are stored; the article is written fresh.
New Event / News Content Package
Provide a brief. The system returns a complete multi-format package — brief, article, Instagram caption, story copy, push notification, video/Reel script, and a poster brief — with a governance verdict and source-grounding check. Then route it to the approval queue.
Paraphrase News into Qatar Living style
Paste a full news article from any source (wire copy, a press release, another outlet). The system re-reports it as an original Qatar Living article — reworded, not copied — applying all 12 Genuine Issues and the house guidelines: clears repeated boilerplate, fixes honorifics, removes sensational language, adds a local-impact angle, suggests a category and byline, and surfaces the key facts. Then it checks how faithful the rewrite is to the source.
Paste the full news article
Compare to a Real Qatar Living Article
Pick a real, recently-published Qatar Living article. The pilot writes its own version from the same verified facts. You judge them side by side.
Rewrite in Qatar Living voice
Paste rough or casual text. LeenAI rewrites it into Qatar Living's institutional register, drops forbidden language, and adds attribution.
Original
Style Compliance Checker
Paste a draft. The checker scores it against seven Qatar-Living style dimensions and lists suggested edits.
Draft
Governance Risk Classifier
Paste any draft or headline. The classifier returns the risk tier, labels detected, and the required Qatar Living approval route.
Draft
Event Freshness Monitor
Surfaces stale events from stale_events.jsonl and proposes a lifecycle action per event.
CMS Cleanup Assistant
Paste a messy article body. The cleaner removes the repeated social-CTA boilerplate, paste artefacts, stray HTML, and whitespace noise (and normalizes markup for SEO/accessibility).
Raw body
Corpus Evidence
Every style rule the AI follows is grounded in the corpus. The chips below come from the actual frequency tables in analysis_data.json.
Top headline openers
The entity-first pattern — 97.3% of QL headlines begin with one of these.
Editorial stance verbs
The verb palette LeenAI prefers. Every one is a top-25 corpus hit.
Avoided language
Probe terms that appear in <1% of the corpus. LeenAI never produces these.
Persona summary
Institutional, formal, optimistic, ceremonial, attribution-heavy. Entity-first headlines. Stance verbs: announced, launched, stressed, affirmed, welcomed, signed. Honorifics respected (29% of articles). Avoids opinion, clickbait, and sensational language.
Sample headlines from the corpus
The house & governance rules
Every draft follows a documented set of house-style and governance rules, derived from this corpus. They are open for your review in full.
Editorial Copilot
The human-in-the-loop queue. AI prepares the draft; the editor approves, edits-and-approves, or rejects with a reason. Every decision feeds the learning loop below — this is how the system improves from editor behaviour over time.
Learning loop
Reviewer reasons aggregated into candidate prompt-rule improvements — the §6 learning hook.
CMS Integration MOCK CONNECTOR
Demonstrates the integration contract, not a live system.
Read-only pull of content opportunities; drafts pushed as
draft status only — never auto-published. A real adapter swaps
these calls for the CMS REST API (e.g. WordPress status=draft).
ROI Calculator
The number executives actually decide on. Conservative by design — adoption is set below 100% and every formula is shown.
Weekly Management Report
The one-page view for management, asked for in the meeting: usage, quality and governance. Numbers marked pilot begin measuring once the controlled pilot runs on real content; the rest are live now.
Operating Model & KPI Readiness
How the system becomes part of the company's daily workflow — every stage has an owner, a KPI, and a control. This is what separates an operating layer from an AI experiment.
Audit trail (every decision logged)
Source · draft · reviewer · decision · reason · timestamp — populated as the approval queue is used.
Qatar Living GI — House & Governance Rules
The rules the system follows, shown in full and open for your review. Nothing is hidden in a prompt. Your team reads each rule, edits it, and approves it; the system updates accordingly. Use the Feedback tab to suggest a change to any rule.
Feedback & Requests
Tell us what to improve, add or change. Write it naturally — the assistant will help shape it into a clear request. Everything you submit is saved and reviewed by the LeenAI team.