Frequently Asked Questions

Sentinel is the first LLM-native narrative intelligence platform built for the MENA region. It monitors how entities - brands, companies, politicians, government institutions - are discussed across Arabic, English, and French media, and converts raw coverage into actionable intelligence including narrative arc detection, momentum tracking, and AI-generated briefings.

Every MENA media intelligence competitor was built before LLMs existed (2015–2022). Their architecture relies on keyword matching and sentence-level sentiment classifiers. Sentinel is built from the ground up on large language models, enabling full-context narrative analysis, cross-language linking, momentum tracking, and contextual sarcasm detection - capabilities that would require competitors to rebuild from scratch.

Sentinel serves PR agencies managing client coverage, brands and corporations protecting their reputation, political campaigns tracking election narratives, and government institutions monitoring public discourse. Our core clients are PR agencies in the MENA region managing multiple client entities.

Sentinel monitors Arabic (MSA + major dialects: Egyptian, Gulf, Levantine), English, and French media simultaneously. Cross-language narrative linking automatically connects the same story across all three languages using multilingual embeddings. The platform UI is available in all three languages with full Arabic RTL support.

Narrative intelligence goes beyond tracking individual articles or sentiment scores. It detects narrative arcs - how a story forms, who drives it, where it peaks, and whether it will resurge. Sentinel tracks narratives through five phases: emergence, growth, peak, decay, and resurgence.

The War Room is a real-time command center for critical moments - debate nights, crises, breaking news. When activated, it switches from the standard 15-minute processing cycle to a 2-minute priority pipeline, providing a live coverage feed with instant multi-channel alerts (email, SMS, Slack). It is designed for rapid response situations where minutes matter.

AI Intelligence Briefings are RAG-powered executive summaries generated by our LLM pipeline with source citations. Unlike dashboard data, briefings are narrative documents that explain what happened, why it matters, and what to expect next. They are generated in your preferred UI language on a daily or weekly schedule.

Traditional tools use sentence-level classifiers that miss sarcasm entirely. Sentinel uses full-article LLM analysis - the model reads the entire article in context, understanding tone, framing, and cultural nuances to correctly identify sarcastic coverage that would otherwise be misclassified as positive.

Visibility is how Sentinel tracks the way the world describes each entity in the places people check first. It has three parts: a canonical profile of each entity (from Wikipedia, Wikidata, and Google Knowledge Graph), a weekly capture of page 1 of Google and Bing, and a weekly read of what ChatGPT, Claude, and Gemini say about the entity, with accuracy flags against the profile. The profile is part of every plan; search presence is available from Intelligence up, and AI-answer visibility from Enterprise up. Every finding comes with a drafted fix.

Social listening and media monitoring tell you what people are saying right now. Visibility answers a different question: how the world describes you in the places decision-makers check first - Google and Bing results, and AI answer engines. Combined with Sentinel's narrative intelligence on news, Telegram, and X, it closes the GEO (Generative Engine Optimization), SEO, and ORM (Online Reputation Management) loop in a single platform.

We offer a 7-day pilot with 1 monitored entity at no cost. This lets you see real intelligence on your actual brand or client - not demo data. During the pilot, all features are enabled so you can experience the full platform. Contact us to start your pilot.

Sentinel offers three plans plus a Custom option: Monitor ($750/mo) for essential monitoring, Intelligence ($3,000/mo) for advanced analysis with spike attribution and dialectal sentiment, and Enterprise ($6K–$8K/mo) for full narrative intelligence with War Room. Political campaigns, sovereign funds, and large multi-entity portfolios are served by a tailored Custom plan. All plans include News, Telegram, and X/Twitter monitoring, and are subscription-based with no long-term contracts required.

Standard subscription plans are month-to-month with no lock-in. The Campaign tier is also available as an election-cycle contract ($55K–$75K per cycle) for political campaigns with defined start and end dates.

Articles are always ingested - we never lose your data. If you hit your processing cap, additional articles are stored and processed when you upgrade. We notify you before you reach limits so there are no surprises.

Available add-ons include extended historical backfill (30 days at $500/entity, 90 days at $1,500/entity, 6 months at $3,000/entity, 1 year at $6,000/entity - covers news, Telegram, X, and Google Trends in one go), Human Analyst Review ($3K–$5K/session), additional monitored entities, and additional War Room hours.

Sentinel monitors three signal layers across the MENA region: 1,000+ curated news outlets in Arabic, English, and French - across 27 countries including Lebanon, Saudi Arabia, UAE, Egypt, Jordan, Qatar, Kuwait, Bahrain, Oman, Morocco, Tunisia, Algeria, and more; real-time social signals from X / Twitter and public Telegram channels; and search-demand data so we know what audiences are actually looking for. Coverage spans major newspapers, news agencies, online outlets, broadcast transcripts, and social posts. We do not scrape or republish full article text - only excerpts, metadata, and source links.

Active subscriptions retain all data for the duration of service. 7 days of historical backfill are included free on all tiers at onboarding. Extended backfill is available as a paid add-on (30, 90 days, 6 months, or 1 year) for deeper historical analysis.

Sentinel is self-hosted with full data sovereignty - your data never leaves your infrastructure. All secrets are encrypted with AES-256. The platform supports per-tenant data isolation, role-based access control, and audit logging. For Enterprise+ clients, a Data Processing Agreement (DPA) is available.

Yes. Sentinel follows GDPR standards which cover all MENA data protection requirements. Users must accept Terms of Service and Privacy Policy at signup. Enterprise+ clients can execute a separate Data Processing Agreement. All data handling follows principle of least privilege with full audit trails.

After activation, real data starts flowing within 2–4 hours. Full onboarding including entity setup, keyword tuning, and alert configuration typically takes 1–2 business days. Your account manager guides you through the process.

Sentinel is deployed via Docker Compose on your infrastructure (self-hosted). The stack includes Next.js, MySQL, Redis, Qdrant (vector search), and Meilisearch. There is no cloud-only option - all data stays on your servers.

The Reputation Score is a composite metric with four components: sentiment (40%), volume & visibility (25%), narrative health (25%), and momentum (10%). Volume factors in prominence-weighted article counts and digital visibility. All component weights are configurable from the admin panel.

Sentinel uses multilingual embeddings to automatically detect when the same narrative appears across Arabic, English, and French media. This means you can see how a story that starts in Arabic press gets picked up in English outlets or French media - all linked to a single narrative thread without manual tagging.

When coverage of an entity exceeds 2x its rolling average, Sentinel automatically identifies: the trigger event that caused the spike, the origin article (first mover), the narrative thread, and related entities. Spike markers appear on timeline charts with clickable analysis cards.

Email and in-app alerts are available on all tiers. SMS alerts are available from the Intelligence tier and above. Slack and webhook integrations are available on Enterprise tier and above. Alert rules can be configured by volume thresholds, sentiment shifts, narrative emergence, or crisis events with configurable cooldown periods.

We don't currently offer a public API. It may be considered in the future on request - talk to your account manager. Today, all interaction is through the dashboard UI, and Enterprise and above clients can use webhook integrations to push intelligence into their existing systems.

All clients have a dedicated account manager. Support is available via email at [email protected]. Enterprise and above tiers receive priority support with faster response times.