Bilingual, in one conversation
Nasaq moves between Arabic and English within the same conversation, dialect included, without the patient ever noticing the switch.
Clinic Operating System
Your clinic doesn't just reply. It moves.
Nasaq connects appointments, patients, follow-up, tasks and the front desk into one operating system for the clinic - turning every interaction into a structured, traceable action.
Zero
conflicting appointments
Every action
logged and traceable
Two languages
in one conversation
Arabic-first · PDPL-aligned
Nasaq
Clinic operating system
Incoming message
Need a cleaning tomorrow after 4
IntentCleaning
PatientSara Ahmed
ServiceCleaning · 30 min
DoctorDr. Ahmed
ChairChair 02
TimeTue 4:30
Conflict checkNo conflict
Appointment confirmed
Reminder scheduledTomorrow · 10:00 AM
Runs the operations layer for modern clinics
DentalDermatologyPhysiotherapyFamily MedicineAestheticsPediatricsOphthalmologyDentalDermatologyPhysiotherapyFamily MedicineAestheticsPediatricsOphthalmology
How Nasaq works
Every message can start an operation.
Whatever the channel, every interaction passes through five fixed steps before it becomes an action inside the clinic.
Patient messageUnderstandVerifyExecuteFollow upEscalate
from any channel
- 01
Understand
Reads the patient's request, dialect and conversation context.
- 02
Verify
Checks the doctor, service, time, resources and clinic state.
- 03
Execute
Books, changes, cancels, or creates a task.
- 04
Follow up
Confirmations, reminders, follow-up and after-visit care.
- 05
Escalate
When it needs a staff member or doctor, it hands off to a human.
The scheduling engine
Booking isn't just picking a time.
Every appointment has to line up the patient, the doctor, the service and the resource at the same moment - with no conflict.
Cleaning30 minutes
DoctorDr. AhmedAvailable
ResourceChair 02Available
PatientSara AhmedAvailable
4:30 – 5:00No conflict
Confirm appointment
Appointment lifecycle
The appointment doesn't end at booking.
Nasaq follows every appointment through to its end - whether it's confirmed, cancelled, or goes unanswered.
The normal path
BookedReminderConfirmed
On cancellation
CancelledSlot releasedWaitlist contactedSlot recovered
On no response
No responseConfirmation follow-upStaff task
Follow-up
The visit ends. The follow-up doesn't.
Patient follow-up runs as a scheduled process, not a task someone has to remember.
1Visit completed
The state changes, and the rest begins
Whitening · completed 14:00
2Follow-up scheduled
Set automatically from the visit
Scheduled automatically, no one has to remember
3Message to patient
In their language, at the right time
Hope you're doing well after yesterday's session. Any pain or sensitivity?
4Patient replies
Read and classified on arrival
A bit of sensitivity in my teeth
5Action or handoff
Closed automatically, or a task for your team
Task created for Dr. Khalid
Human + system
Automation where it can. A human where it must.
Nasaq executes what's safe and defined on its own. Anything that needs human judgment is escalated immediately, with full context preserved.
Discount requestTask for front desk
X-ray imageTask for the doctor
Sensitive caseDirect handoff to staff
- Creates a task for the team
- Preserves the full conversation context
- Records the action in the log
The architecture
Two layers, one hard boundary
Nasaq is split on purpose. A thin AI interface that understands language sits on top of a deterministic engine that owns every decision and every write — and the two can only ever talk through a guarded gate.
Intelligent interface
proposesReads what the patient means in Arabic or English, asks for anything missing, and phrases the reply. It suggests — it never writes.
- Understands intent
- Speaks two languages
- Drafts every reply
Guarded interface
Reliable operating system
decidesRe-checks live availability and clinic rules, then writes the appointment through one guarded path with tenant isolation and a conflict constraint.
- Validates reality
- Enforces the rules
- Owns every write
Intelligence never decides alone. Every action passes through the clinic's rules.
It can understand a request, propose real open slots, and draft the confirmation.
It can't invent a slot, bypass a rule, or make a single clinical call.
System of record
The patient's journey, not just the chat log.
Nasaq doesn't just remember the last message - it owns the full operational state of every patient.
SSara Ahmed
- Jun 2Asked about a cleaning
- Jun 2Appointment booked
- Jun 3Reminder sent
- Jun 4Visit completed
- Jun 5Follow-up sent
- Dec 4Recall appointment scheduled
What Nasaq does
An operations layer, not another chatbot.
Every capability below rests on one rule: a deterministic core decides, and the AI only understands and talks. That is what makes it safe to put in front of real patients.
Just a chatbot
Guesses, hallucinates, double-books
A clinic operating system
Understands, verifies, executes, and never double-books
One dashboard for the desk
Your front desk sees every conversation, appointment and task in a single calm view - no app-switching, no copy-paste between tools.
WhatsApp is the front door, not the system
A patient starts on WhatsApp today, and on the web or another channel tomorrow. Nasaq owns what happens after any message, from any channel.
Fewer no-shows. Zero double-bookings. Calmer front desk.
- Appointments without conflicts
- Follow-up that doesn't rely on memory
- Every action is traceable
- The schedule stays in sync
Safety & trust
Designed so the AI can't do the dangerous thing
In healthcare, an unpredictable AI is a liability. Nasaq's guarantees are structural - enforced by the database and the architecture, not by hoping the model behaves.
No clinical decisions
No diagnosis, treatment or drug advice on any surface. The model handles scheduling and language only; anything clinical escalates to a human by rule.
Consent before any data leaves
No patient information reaches an external model before explicit, recorded consent. The consent gate is enforced deterministically, not left to the conversation.
Strict tenant isolation
Every clinic's data is walled off at the database level. A cross-clinic query simply cannot be written - isolation is enforced by the database, not a forgotten filter.
Everything is audited
Every state change writes an append-only record of who did what, when, and what changed. The trail is never edited or deleted.
Questions
The things clinics ask us first
No - and it can't. Nasaq draws a hard line: the AI only understands language and schedules appointments. Anything clinical - diagnosis, treatment, medication - is escalated to your staff by an explicit rule, never answered by the model.
Overlapping appointments on the same doctor or resource are made impossible by a constraint in the database, not by an app-level check that could be bypassed. The AI physically cannot promise a slot that isn't really free.
Each clinic's data is isolated at the database level, so a cross-clinic query can't happen. No patient information reaches an external model before recorded consent, and every change is written to an append-only audit trail. The design is aligned with PDPL.
Yes - Arabic-first, including dialect. Patients can write naturally in Arabic or English, and Nasaq replies in the same language. The dashboard is fully bilingual and right-to-left aware.
WhatsApp for patients, and a clean web dashboard for your team. No app for patients to install - they just message the clinic like they already do.
Days, not months. We connect your calendar, services and working hours, tune the reminders, and your clinic has an operating system running from the first message. Your team keeps working the way it already does.
Run your clinic as one system, not scattered tools.
See Nasaq handle a real booking, end to end, from message to execution, in your own clinic's setup. No card, no commitment.
Set up in days, not months. Arabic and English from day one.