Arabic Voice AI Agents: Every Call Answered, Every Booking Captured

It's 7:40 p.m. on a Tuesday in a two-chair dental clinic in Jeddah. Both receptionists left at six. The evening doctor is with a patient. The phone rings four times between 7:30 and 8:15 — a mother asking if the clinic takes walk-ins tonight, a patient wanting to move tomorrow's cleaning, two new patients who found the clinic on Google Maps. All four hit a voicemail nobody will listen to. Two of the four call the clinic two streets over, which answered. On Friday, the clinic is closed and the phone rings eleven more times.

The owner never sees any of this — a missed call generates no report, no ticket, no line item. The loss is real but invisible. As arithmetic, not a claimed statistic: a clinic receiving 40 calls a day that misses a quarter of them at peak hours and after hours misses roughly 300 calls a month. If only 12% of those were bookable visits at an average ticket of SAR 250, that is roughly SAR 9,000 a month walking to competitors — more than 11 times the SAR 800/month line fee. The problem was never rudeness or understaffing. It's that missed calls are invisible because nothing measures them.

Every month you see calls answered, completion rate, appointments booked, and after-hours recoveries — in the exact format of your day-30 missed-call recovery report: "214 calls answered that you would have missed; 37 appointments booked." That is the report template shown with clearly labeled sample data, not a client result — until a real pilot produces real numbers, we never present it as one.

Assess (2–3 days): call-volume review, number and forwarding setup, calendar access, dialect selection. Pilot (~1 week): the agent is configured on your booking rules and benched against our clinic-call dialect evaluation set before it takes a single real call. Prove (30 days): the agent runs on real calls, fully instrumented, with weekly stats and a day-30 missed-call recovery report ending in a go/no-go call. Scale (2–4 weeks): more lines, branches, and reminder campaigns, deeper calendar or CRM integration, staff training, and a runbook. Operate (ongoing): the monthly per-line fee — monitoring, dialect tuning, telephony operations, and the monthly report. Stage 3 is where most AI pilots die; ours ends with a report and a decision, not a drift.

The Voice Agent Pilot is SAR 1,500 setup plus SAR 800/month per line: one week of setup, then 30 fully instrumented days covering booking, reminders, and routing in one dialect, calendar hookup, WhatsApp confirmations, weekly stats, and the day-30 recovery report with a go/no-go call. Cancel anytime during the pilot. After the pilot, the ongoing price is from SAR 800/month per line, which includes dialect tuning, telephony operations, monitoring, and your monthly report — the monthly fee is the retainer, with nothing hidden on top except your telecom line charges. Multi-line clinics and chains get volume pricing on request. Enterprise call-center lines are scoped and quoted individually. Egyptian clinics are served on a separate EGP pricing band on request. Exclusions: no diagnosis or triage advice, ever; more than one location and custom EHR integration are quoted separately.

Don't take our word for it — call the demo line. Our agent will answer in Arabic, book a test appointment, and send you a WhatsApp confirmation. We also publish three recorded demo calls (labeled as our own internal test calls) handling a booking, a reschedule, and a human handoff in Gulf and Egyptian Arabic, plus a summary of our dialect evaluation methodology: a 50-call clinic-audio evaluation set we bench multiple Arabic speech providers against before every go-live, with a sample evaluation report and a sample pilot report (illustrative data) shown on the page. We never invent client names, logos, or results — where we don't yet have client proof, we show you the methodology instead.

Clinics are our first vertical because missed calls translate directly into missed bookings — the clearest ROI in voice AI. The same agent covers overflow and after-hours calls for real estate agencies, F&B groups, and other service businesses. The Middle East contact-center-as-a-service market is projected to more than double, from $479M in 2025 to $1.12B by 2032 (Fortune Business Insights), while enterprise vendors largely ignore mid-market lines. Enterprise deployments ride the same eval and monitoring discipline as the clinic product, with custom integration scope and pricing.

Telephony: local numbers and call forwarding in KSA, UAE, and Egypt through licensed telecom partners — you keep your existing number. Booking and records: Google Calendar out of the box, plus clinic management and booking systems that expose an API; custom EHR integration is scoped separately as fixed-price work so the pilot is never delayed. Speech layer: we bench multiple Arabic ASR and realtime speech providers — including Intella and Munsit — on our own clinic-call evaluation set and deploy the best performer per dialect, re-benching quarterly as the model layer keeps improving.