What a WhatsApp chatbot costs in 2026, broken down honestly

A WhatsApp chatbot has four cost lines: the build, a platform subscription to a Business Solution Provider, Meta's per-conversation charges, and ongoing maintenance. A straightforward service bot typically costs 4,000 to 12,000 US dollars to build, with platform fees from 50 to 500 a month and conversation charges that scale directly with volume. Quotes that give one number are hiding at least two of these.

What drives the build number

Integration is the variable that matters. A bot answering from a fixed set of responses is a few days of work. A bot that checks a live order status, verifies a customer against your CRM, takes a payment and writes the result back to your ERP is a systems project that happens to have a chat interface. When comparing quotes, compare the integration list rather than the number of conversational flows.

Language is the second driver. Arabic support, particularly handling the dialects customers actually type rather than Modern Standard Arabic, adds genuine work: intent coverage has to be built and tested per dialect, and the testing needs native speakers. Treat bilingual as roughly a third more than monolingual, not double.

The conversation charges people forget

Meta bills per conversation by category, and marketing conversations cost several times service ones. A campaign sending a promotional template to fifty thousand contacts has a real and immediate cost that a service bot answering inbound questions does not. Model this before committing to a broadcast strategy, because it is the line that turns a modest monthly bill into a large one without warning.

Maintenance is not optional

Budget fifteen to twenty-five per cent of build cost annually. It goes on reviewing conversations that failed, extending intent coverage as customers ask new things, updating templates when Meta's policies change, and adjusting when your own products and processes change. A bot that is never tuned degrades quietly: customers find the gaps, stop trusting it, and route around it to a phone number.

How to keep the first project small

Pick the single highest-volume repetitive question your team answers, and automate only that, end to end, with a clean handover to a human for everything else. This is usually order or delivery status. It is cheap to build, it is measurable within a fortnight, and it produces the conversation data that tells you what to build second. Attempting full coverage on the first attempt is the most reliable way to spend the whole budget on flows nobody uses.