What is a WhatsApp chatbot and how do you set one up for your business?
A WhatsApp chatbot is software that answers the messages customers send you on WhatsApp automatically. The main setup decision is which kind of account the bot connects to: the WhatsApp Business app on a phone, or the WhatsApp Business Platform (Cloud API), which software can talk to. For a bot that reads from or writes to your booking, order or customer systems, you will in practice need the second.
The WhatsApp Business app is a phone app designed for small businesses to chat with customers one to one; it offers simple automatic replies such as a greeting and an away message. The WhatsApp Business Platform is an API for developers. When a customer writes, the message reaches your server through a webhook and your own software sends the reply. That is what allows the bot to work with your other systems.
What sits inside the bot comes in two forms. A rule-based bot follows menus and fixed flows: "1 for appointments, 2 for order status". An AI-assisted bot understands free-form sentences and answers from your company's documents. Turkish searches around this topic include phrases such as "whatsapp chatbot oluşturma" (building a WhatsApp chatbot) and "whatsapp chatbot n8n"; both come back to the same choice of account type and bot design.
Which setup path suits your business?
The table below compares four paths. The choice depends less on daily message volume than on whether the bot has to connect to other systems.
| Path | When it fits | What to watch |
|---|---|---|
| WhatsApp Business app | Few messages, one person replies | No system integration or real bot |
| Platform + ready-made bot tool | FAQs and simple flows are enough | Monthly subscription, limited flexibility |
| Platform + custom software | Booking, order or CRM connection | Needs development and upkeep |
| Platform + AI assistant | Open questions, wide knowledge base | Risk of wrong answers, handover essential |
For most businesses the right start is a narrow bot that does one job well. Adding an AI layer later is less risky than trying to do everything from day one.
What are the cost lines of a WhatsApp chatbot?
The cost is more than a single "WhatsApp price". Meta's official pricing page (as updated on 30 September 2026) describes per-message pricing: you are charged only for delivered messages from the business to the user, and messages from users to the business are not charged. The rate depends on the message category (marketing, utility, authentication, service) and on the market of the recipient's country calling code. Rates change often, so this article gives no fixed amounts; check the page for the current rate.
The same page describes two free cases. First, each business phone number gets its first 1,000 service messages each month free, and unused allowance does not roll over. Second, when a user messages you from a click-to-WhatsApp ad, the free entry point window that opens makes messages free; it can stay open for up to 7 days. According to the page, charging for service messages resumed on 1 October 2026, so the monthly 1,000 allowance now matters in a budget.
| Cost line | Paid to | Note |
|---|---|---|
| Message fees | Meta | Per delivered business message |
| Bot software or platform subscription | Software provider | Monthly or usage-based |
| Initial setup and integration | Software company | Booking, CRM, order system links |
| AI model usage | Model provider | Only for AI-assisted bots |
| Maintenance and content updates | You or your supplier | Prices, hours, campaigns change |
| Live support staff | Your own team | Someone to take over when the bot cannot |
For Meta's own AI agent the pricing page sets out a separate model: USD 2.00 per million tokens. A message typically uses 20,000 to 25,000 tokens, which comes to roughly 4 to 5 US cents per message. This is the value published at the time of writing and it may change.
How do you set up a WhatsApp chatbot step by step?
- Pick one job. Appointment reminders, order status or frequently asked questions: something whose result you can measure.
- Make your company the account owner. Open the business account and phone number in your own name, so you keep access even if a software company does the setup.
- Choose the access route. Either the Cloud API directly or through a business partner. In both cases incoming messages reach your software by webhook.
- Prepare message templates. The WhatsApp business policy says you may start a conversation only with an approved template. Within the 24-hour window after a user writes, you can reply freely without one.
- Write the conversation flow. Decide what the bot does whenever it does not understand.
- Build the handover to a person. The same policy asks for a prompt, clear escalation path when you use automation, such as in-chat transfer to a human agent, a phone number, email or a support form.
- Test, then open gradually. Try it with staff first, then a few customers, and watch answer accuracy and the handover rate.
How should customer consent and KVKK be handled?
Under the WhatsApp policy you may contact a person only if they have given you their number and have opted in to receive later messages from you. Choosing a lawful opt-in method is the business's responsibility. If someone blocks you or asks to stop, you must respect that.
Two further Turkish laws apply. Marketing and campaign messages count as commercial electronic messages, and Article 6 of Law No. 6563 says they may be sent to recipients only with their prior consent. The same article contains an exception for tradespeople and merchants, whose scope should be checked against current legislation. For personal data, Article 10 of the KVKK (Law No. 6698) requires the data controller to tell people who it is, why the data is processed, to whom it may be transferred, how it is collected and on what legal basis, and what rights the person has. The bot's first message is a sensible place to link to a short privacy notice.
If message content is processed by a provider abroad or by an AI model, further assessment may be needed; we cover this in our guide to KVKK and generative AI. We recommend confirming with your legal adviser which form of consent is sufficient for you.
Which jobs suit a WhatsApp chatbot?
- Booking and reminders: The customer picks a date, the bot writes it to the calendar and reminds them the day before.
- Order and delivery status: The customer asks with an order number; the answer comes from your system, not from the bot's imagination.
- Frequently asked questions: Opening hours, address, returns policy and other fixed facts.
- First-contact details: The bot collects a name, a need and a phone number and passes them to the sales team.
Do not leave complaint handling, contract interpretation or emotionally charged conversations to a bot. We discuss this more widely in our article on AI in customer service.
How many messages will you send, and how many are free?
For example, suppose a dental clinic receives 600 appointment requests a month. For each request the bot sends an average of 3 replies: date options, a confirmation and some information. On that assumption it sends 1,800 business messages a month. If all of them go out within the 24-hour window after the patient writes and count as service messages, the first 1,000 are free and the remaining 800 are charged at Meta's published rate.
If the clinic also sends a reminder the day before, that message goes outside the window as a template and is charged in whichever category Meta's template classification assigns it. If you run the same calculation for an AI-assisted bot, model usage is added as a separate line. The figures are an illustration only; the real amount depends on your number, the country market and the message categories.
The method that follows is simple: estimate how many business messages you will send each month, split them by category and multiply by the current rate card.
Common mistakes when setting up a WhatsApp chatbot
- Sending campaign messages without consent, which can lead to restrictions on the account under the policy and to legal exposure.
- Letting an AI bot answer freely with no source material, which risks wrong prices or promises. We explain the causes in our article on AI hallucination.
- Forgetting the handover to a person, when the policy expects a clear escalation path alongside automation.
- Budgeting by number of users, when the number and category of messages decide the fee.
- Opening the business account in the software company's name, so you may lose the number and history if you part ways.
- Writing the bot's answers once and leaving them, so it starts giving wrong information as prices, hours and campaigns change.
How Hatson Tech helps
Hatson Tech builds WhatsApp bots that connect to your booking, order and customer systems as part of its AI agent development service. The first meeting and discovery are free: together we work out which job the bot should do, which conversations go to a person and what your monthly message volume looks like. Hatson does not publish fixed prices; once the scope is clear we prepare a proposal for you. If you do not want data to leave your organisation, we can run models on your own servers. You can read about approval points between a bot and a person in our article on human approval in agents.