Arabic Customer Messaging: Best Practices for GCC Brands

This guide is for CX, marketing and operations teams that message customers in Arabic across WhatsApp, SMS, email and social inboxes. It covers what to write, how Arabic behaves technically on phones, and how agents should work in two languages, so you can set a style guide and a QA checklist for your team.

Which Arabic should customer messages use: MSA or Gulf dialect?

Use Modern Standard Arabic (MSA, often called fusha) for anything templated, and let agents adapt to dialect in live conversations. MSA is understood across the GCC and the wider Arab world, reads as professional, and is safer for amounts, policies and legal wording. Gulf dialect (Khaleeji) feels warmer but varies between countries and can seem informal or even off-brand for banks, government-linked services and healthcare.

Many GCC customers are not Arabic speakers. A large share of residents in the UAE and Qatar especially are expatriates, so Arabic should be one of your languages, not an assumption.

How do you build bilingual Arabic and English templates?

Create two templates, not one mixed message. WhatsApp templates are approved per language, and Meta lists Arabic as a supported template language, so you can keep an Arabic and an English version under the same template name and send whichever matches the customer's stored preference.

Template writing procedure

1. Write the Arabic first, natively, rather than translating English word for word. Arabic sentence structure and politeness formulas differ.
2. Keep variables meaningful: a greeting with the customer's name, a date, an amount. Avoid stacking several variables in one line.
3. Match the meaning of both versions so that support agents answering either language see the same commitment.
4. Mind the length. Arabic can be shorter or longer than English depending on phrasing; check how the message looks on a small screen.
5. Use sample values in review that look like real data, including long names and long order numbers.
6. Get a native reviewer to sign off greetings, honorifics and tone before submission.
7. Store language preference per contact, and default to bilingual sends only when it is unknown.

For template categories and approval rules, see WhatsApp template categories and approval.

How does right-to-left and mixed-direction text behave?

Phones render mixed Arabic and Latin text with the Unicode Bidirectional Algorithm (UAX #9). Arabic runs right to left, while Latin words and digits run left to right inside it. The algorithm decides where "neutral" characters such as spaces, slashes, hyphens and punctuation go, and that is where most display bugs appear.

Common symptoms and fixes:

Unicode also defines directional control characters, such as the right-to-left mark and isolate characters, and the W3C recommends isolates where markup is not available. Some channels and phones handle them inconsistently, so test before relying on them. In email, use the HTML dir attribute on the message body.

Which digits, dates and currencies should you use?

Pick one convention per message type and document it. GCC customers read both Western digits (0-9) and Eastern Arabic digits, but mixing them looks careless and can break copy-paste.

• OTP codes, order and account numbers, phone numbers: Western digits, so codes paste correctly into apps.
• Prices: Western digits are common in GCC retail; include the currency explicitly, for example AED, SAR or QAR, in the language of the message.
• Dates: spell the month name ("15 مارس" or "March 15") rather than writing 03/15, which customers may read in either order. If you use Hijri dates, show the Gregorian date as well for operational messages.
• Times: state the time zone when serving more than one GCC country.

How should you handle Arabic names?

Store names exactly as the customer provides them, in the script they used. Many customers will have an Arabic-script name in one system and a Latin transliteration in another, with different spellings across documents (Mohammed, Muhammad, Mohamed). Do not "correct" spellings automatically.

• Keep separate fields for Arabic and Latin names if your CRM supports it.
• Use the given name in greetings; family structures and naming conventions vary, so avoid assuming a surname.
• Check that CSV import and export preserve Arabic characters; save files as UTF-8.
• When the name is missing, use a neutral greeting rather than a placeholder that renders empty.

Why do Arabic SMS cost more segments?

Arabic SMS uses UCS-2 encoding. Under 3GPP TS 23.038, a single SMS fits 160 characters in the GSM-7 alphabet but only 70 in UCS-2, and multi-part messages lose a few characters per segment to headers. One Arabic character, emoji or certain symbols switches the whole message to UCS-2.

That means a 200-character Arabic notification is usually three or more segments and is billed accordingly. Write short Arabic SMS, avoid emoji, and move longer content to WhatsApp or email. Sender ID rules in the UAE and Saudi Arabia apply equally to Arabic messages; see SMS sender ID registration.

How should agents work in Arabic and English?

Give agents rules, not just a translated interface. Your shared inbox should route by language where possible and support right-to-left replies.

Agent workflow checklist

1. Detect or store each customer's language and route to an agent who reads it.
2. Reply in the customer's language; switch only if the customer switches.
3. Keep saved replies in both languages with identical commitments.
4. Escalate to a native speaker when dialect, slang or a complaint is unclear.
5. Confirm key facts (amounts, dates, references) in plain MSA even in a dialect chat.
6. Never paste machine translation into a complaint or legal answer without review.
7. Tag conversations by language so reporting shows demand per language.

Worked example: a Doha clinic group

"Sidra Care Clinics" (synthetic) sends appointment reminders and handles booking chats on WhatsApp in Arabic and English.

• Templates: an Arabic and an English reminder under one template name, with patient first name, clinic branch, date written with the month name, and time with "Doha time".
• Digits: Western digits for times and booking references, because patients copy references into the booking app.
• Direction: the booking reference sits on its own line to stop it jumping position in the Arabic version.
• Agents: four bilingual agents, one Arabic-only senior agent for escalations; saved replies for rescheduling in both languages.
• QA: every template tested on iOS and Android with a 30-character name and a 12-digit reference before submission.

What mistakes should you avoid?

• Translating English templates literally into Arabic.
• Using dialect for financial, medical or legal content.
• Testing with short placeholders that hide direction bugs.
• Mixing Western and Eastern Arabic digits in one number.
• Exporting contacts in a non-UTF-8 format and corrupting names.
• Sending long Arabic SMS without checking segment counts.
• Assuming every GCC customer prefers Arabic.

How OmniChannels.ai helps

OmniChannels.ai supports Arabic and English across the product. You can create WhatsApp templates with variables, media and buttons, preview them on a phone mock-up and track approval. Campaigns accept Excel or CSV uploads for WhatsApp, SMS, email and Telegram with preview and duplicate filtering, and the Omni Inbox brings WhatsApp, Instagram, Telegram and email conversations into one place, with templates and attachments in replies. Delivery reports with filters and CSV export help you review results by language segment.