How Much Does an AI WhatsApp Chatbot Cost for a Business in Saudi Arabia in 2026?

The first kind is a menu bot: “Press 1 for prices, 2 for branches.” It is built on a subscription platform in days, it is cheap, and it is enough if your customers’ questions are limited and repetitive. Its weakness is that a customer who types their question in their own words gets no answer, so they move to a staff member or leave.

The second kind is an AI bot that understands Arabic and the Gulf dialect and answers from your company’s information: prices, services and policies. The third is an agent that works on your behalf: it books an appointment in your system, sends a quote, checks an order’s status, logs the customer in your CRM and hands them to a sales rep. The price difference between the three reflects the difference in the work the bot actually does.

These are the elements that move the cost in WhatsApp projects; the clearer you are about them in your request, the more accurate the quote:

The official WhatsApp Business API charges for messages the business starts: in Saudi Arabia, about SAR 0.19 for a marketing message and about SAR 0.04 for a utility message or verification code, according to prices published by service providers. Replies within a conversation the customer started are treated differently, so the bill depends on how you use it.

An AI bot adds the cost of the language model per conversation, hosting, and ongoing monitoring to improve its answers. Market estimates put running costs for integrated automation at SAR 3,000–8,000 a month. Compare that with the salary of a customer service employee who works limited hours, while the bot replies around the clock.

The most common is using a personal number or the regular WhatsApp app with unofficial tools; the number gets banned after a bulk send and the conversation history is lost. Use the official WhatsApp Business API from the start, and ramp up sending gradually from a new number.

The second mistake is letting the bot give wrong prices or commitments because it was not given accurate information or clear limits. A good bot knows when to say “let me pass you to a colleague” and hands the conversation over with a summary. The third is launching without reviewing its conversations in the first weeks to improve its answers.

Ask for a live test on real questions from your customers, in Arabic and in dialect, then compare:

Start with the ten questions you get most every day and one task that eats your team’s time, such as booking an appointment or sending a quote. Launch the bot on those, review its conversations weekly, then add tasks gradually. This is cheaper and more successful than building a bot that tries to do everything from day one.

At Fortune Code we build WhatsApp agents that understand Arabic and connect to your systems, with a written scope and a contract before we start, and payments tied to delivery. We work on Saudi time with daily follow-up on WhatsApp, and send a written proposal within 24 hours.

Work out the return with three questions. First: how many messages do you get a day, and how many are repeat questions about prices, hours and branches? Second: how many leads are lost because the reply came late, after hours or at the weekend? Third: how many hours does your team spend copying and pasting the same answers?

If the bot will answer half of the repeat messages, catch the customers who message at night, and book or quote them before they go to a competitor, the return usually shows within a few months. If your messages are few and every conversation involves its own negotiation, an organised shared inbox on a simple subscription platform may be enough to start with.

A bot is only as good as the information it is built on. Prepare these before you start to shorten the timeline and get accurate answers from day one:

Ready-made WhatsApp platforms are excellent for organising a shared inbox across several staff, sending campaigns and menu-based replies, at a low cost. But they are generic by nature, and know nothing about your quotation, booking or inventory system without extra integrations.

A custom agent is built around your business: it reads prices from your system, books into your calendar, logs the customer in your CRM, and knows when to hand over the conversation and to whom. It costs more upfront, but it does work that used to need an employee. Many companies start with a ready-made platform and move to an agent once they know which tasks consume their team’s time.

Watch four numbers in the first month: the share of conversations the bot completed without staff, the average first response time, the number of bookings or quotes it created, and the conversations handed to staff and why. The last one is your guide to improving the bot, because it reveals the questions it has not yet been trained on.