Sanafi · سنافي · Consular early warning
Know within minutes who needs you, wherever they are.
Sanafi is an agentic AI platform for consular early warning and citizen protection. AI agents watch approved sources around the clock, score every event, match it against registered citizens and prepare messages in Arabic and English. An officer decides every step.
- Eight AI agents, each with clear limits
- Four gates where officers decide
- On premises or sovereign cloud, no external AI calls
Sanafi · Operations Centre
Live · every 2 minGate 2 · Approve the message
Severity 1Welfare check for 14 registered citizens near the rail incident
نطمئن على سلامتكم. يرجى الرد بالرقم 1 إذا كنتم بخير.
We are checking on your safety. Please reply 1 if you are OK.
14 people · SMS, app, instant messaging and email · approver: duty supervisor
Authorized vendor to
- Ministry of Finance
- Ministry of Interior
- DGE Abu Dhabi
- Etihad Aviation
- EDGE Group
Delivered for
- Hindustan Unilever
- Nokia
- Linde
- USA Today
- JSL
Procurement-ready
- Abu Dhabi Government Procurement Gate
- UAE Ministry of Finance, federal procurement
- Registered ADNOC vendor
- National ICV certificate
Product tour
See Sanafi at work
Real screens from the Sanafi Operations Centre, shown with sample data.
Agents prepare, officers decide
Watch, scoring, impact and alerting agents do the groundwork. Nothing goes out until an officer approves it at a gate.
Every step on the record
Each agent action and each officer decision is logged in UAE time, ready for review and audit.
Arabic and English
The whole Operations Centre switches between Arabic and English.
Why Sanafi
The first hour matters most
When an incident happens abroad, information reaches the ministry through separate routes: news, partners, missions and calls from citizens. Joining it up takes the time that matters most. Sanafi does it in minutes.
Signals from many sources
Official agencies, international bodies and established news, in Arabic and English, merged into one event with its sources.
Who is affected
Sanafi draws the affected area and counts the registered citizens in it, showing the basis and age of every match.
Messages in two languages
Arabic and English drafts from approved templates, sent only after an officer approves.
AI agents
Eight agents, each with clear limits
Every agent works within limits the ministry sets. Each action is logged, and anything that reaches a citizen or settles a decision waits for an officer.
Watch agent
Reads every approved source around the clock, in Arabic and English, and merges reports of the same event.
Scoring agent
Grades authenticity and severity with the reason for every point, and rescores when the facts change.
Impact agent
Draws each affected area and matches the registered citizens in it.
Alerting agent
Alerts the duty officer and the missions in the country, and escalates when an acknowledgement is late.
Messaging agent
Drafts citizen messages from approved templates in Arabic and English, and checks recipients and duplicates.
Follow-up agent
Reads every reply, opens cases in priority order and queues calls to anyone who has not replied.
Briefing and learning agent
Drafts situation briefs and after-action reports, finds similar past incidents and proposes lessons.
Service watch agent
Checks the ministry’s digital services every minute and opens a ticket when one fails.
How it works
One loop, with four gates for officers
The agents watch, score and prepare by themselves every two minutes. From the first gate on, nothing moves until an officer decides.
- 01
Confirm the incident
The duty officer confirms, keeps or dismisses each incident, with a reason.
- 02
Approve each message
An officer approves every citizen message. Severity 1 messages go to the duty supervisor.
- 03
Escalate or invoke crisis mode
When a situation needs more than the duty team, the duty supervisor decides.
- 04
Close
The duty officer closes the incident when every matched person is accounted for.
Transparent scoring
Every score shows its reasons
The score card shows every point and why it is there. Officers accept or change each score, with a reason.
Authenticity
Every source has a reliability grade from A to F. Independent signals are combined, with caps that stop overconfidence.
Severity
Harm is weighed against the number of people exposed, with modifiers for context and clear minimum and maximum rules.
Route
Severity and grade together set the route: alert, verify, watch or the daily summary.
Operations Centre
Everything the duty team needs, in one place
Ask Sanafi, the duty assistant
Officers ask in Arabic or English, such as “Which alerts are waiting for an officer?”. It answers from the live picture with read-only tools and takes no decisions.
People abroad
Registered citizens and app check-ins: who is inside an affected area now and who is due to fly in the next 24 or 72 hours. Only people who registered are counted.
Messages on every channel
Approved messages can go out by SMS, the ministry’s app, instant messaging and email, with Arabic and English side by side.
Missions and contact centre
Missions confirm local facts, and the contact centre gives callers the current advice and opens their cases.
Crisis mode
The duty supervisor brings in the crisis team, briefs the head of mission and extends contact centre hours, following the ministry’s protocol step by step.
Exercise mode and lessons
Run training exercises in exercise mode, and keep every incident’s history, after-action report and lessons.
Use cases
Built for the events consular teams face
Transport incidents
Rail, air and maritime incidents on the routes and in the cities your citizens use.
Security incidents and unrest
Attacks, protests and unrest, confirmed across sources before anyone is alerted.
Natural hazards and severe weather
Storms, floods, earthquakes and wildfires, from the warning onwards, matched to the citizens in the area.
Planned disruption and travel advice
Strikes, closures and major events flagged days ahead, and changes to official travel advice as they happen.
Track record
Trusted by UAE government entities and global brands.
Authorized vendor to five UAE government entities, with projects delivered for Hindustan Unilever, USA Today, Nokia and Linde.
- Startups and enterprises served by our engineering teams
- 150+
- Startups and enterprises served by our engineering teams
- Average rating on Clutch, Trustpilot and Glassdoor
- 4.9/5
- Average rating on Clutch, Trustpilot and Glassdoor
- Client satisfaction rate
- 98%
- Client satisfaction rate
- UAE government entities we are an authorized vendor to
- 5
- UAE government entities we are an authorized vendor to
Hear it from our clients
Client testimonial
E-commerceClient testimonial
HealthcareClient testimonial
E-commerceClient testimonial
E-commerce
Governance
Sovereign by design
Inside your perimeter
Data, processing and AI models stay on the ministry’s premises or in a sovereign cloud of its choice, with no external AI calls.
The officer decides
No message reaches a citizen without an officer’s approval. Severity 1 messages are approved by the duty supervisor.
Every action logged
A full audit trail of every agent action and officer decision, with a monthly review of model performance.
Drafts, never decisions
Each agent’s language model can only draft or explain. It cannot send, decide or change anything.
Book a demo
See Sanafi work on today’s events
A product specialist will walk you through the live Operations Centre on real events from public sources, with a sample registry, in Arabic or English.
- The live Operations Centre, not slides
- Real events from today
- A reply within one business day
4.9/5 average rating · 98% client satisfaction · 150+ clients served
FAQ
Questions buyers ask
What is Sanafi?
Sanafi is an agentic AI platform for consular early warning and citizen protection, built for foreign ministries. AI agents watch approved sources around the clock, score every event, match it against registered citizens and prepare messages in Arabic and English. An officer decides every step.
Does Sanafi send messages on its own?
No. No message reaches a citizen without an officer’s approval, and Severity 1 messages are approved by the duty supervisor. The agents watch, score and prepare by themselves; from the first gate on, nothing moves until an officer decides.
How does Sanafi score events?
Every source has a reliability grade from A to F, and severity weighs harm against the number of people exposed. The score card shows every point and why it is there, and an officer accepts or changes each score with a reason.
Which sources does Sanafi read?
The sources the ministry approves: official agencies, international bodies, national news outlets and news agencies, Google News and social media, in Arabic and English. Every report carries the reliability grade of the outlet that published it.
How often does Sanafi check for new events?
Every two minutes. In each cycle the agents read the sources, merge reports of the same event, score it and match it against registered citizens.
Where does Sanafi run?
On the ministry’s premises or in a sovereign cloud of its choice. In production every agent runs on the ministry’s own model, with no external AI calls.
Can other organizations use Sanafi for travel risk and duty of care?
Sanafi is built for foreign ministries. The same approach can serve any organization that looks after people abroad, such as staff on assignment or business travel. Tell us about your case and we will show you what fits.
How can a ministry procure Sanafi?
Federal ministries and entities can buy through the UAE Ministry of Finance’s Digital Procurement Platform, where Birbal AI is a registered supplier, and Abu Dhabi government entities through the Abu Dhabi Government Procurement Gate. Birbal AI also holds a National In-Country Value (ICV) certificate.
Know who needs you, within minutes
See Sanafi work on today’s events in a live demo.
Book a Sanafi demo