DEV Community: FIRELIFE Ventures

WhatsApp Business API, explained: how it works and what decides the cost

FIRELIFE Ventures — Sat, 14 Nov 2026 19:25:10 +0000

With the WhatsApp Business API, Meta charges for each delivered message your business starts, at a rate set by the message's category and the country of the customer's number. Replies to customers who wrote first cost far less, so the cheapest WhatsApp set-up is one where customers start the conversation and your updates are written as utility messages.

In short

The API connects your WhatsApp number to your own software; Meta charges per delivered message, never for messages customers send you.
Messages you start must be approved templates: marketing, utility or authentication, each at its own rate per country.
When a customer writes first, a 24-hour window opens for replies: since 1 October 2026 the first 1,000 service replies a month per number are free.
Your monthly bill follows how many messages of each category you send, and to which countries; the build is a separate, one-time cost.
Start with utility messages customers want, such as order updates, and get opt-in before any marketing.

What is the WhatsApp Business API, and who charges for it?

The WhatsApp Business API connects your WhatsApp number to your own software, and Meta charges for the messages your business sends through it.

The free WhatsApp Business app runs on a phone. One or two people answer chats by hand, and whatever they type stays on that phone. The API is the other way in: your number is linked to your online store, CRM, booking system or ERP, so messages go out by themselves, a whole team can answer from a shared inbox, and every chat sits next to the right customer and order.

There are two kinds of cost, and it helps to keep them apart from the start:

Meta's message charges. Meta sets the rates and bills per delivered message. These are the same whichever company helps you set things up.
The build and the running of it. Somebody connects WhatsApp to your systems, writes the templates and builds the inbox, flows or AI assistant. That is a one-time cost, plus care afterwards. Some providers also charge a monthly platform fee on top of Meta's rates.

This guide is about the first kind: how Meta's pricing works, and how to estimate what a month will cost you.

How does WhatsApp Business API pricing work?

Meta charges for each delivered message your business sends, at a rate set by the message's category and the country of the customer's phone number.

Since 1 July 2025, pricing has been per message. Before that, Meta charged per 24-hour "conversation", which is why older articles still talk about conversation fees. Today, every message has exactly one category and is charged at most once, and only when it is delivered. Messages customers send to you are never charged.

Two things decide the price of one message:

Its category. Marketing, utility, authentication or service (more on each below).
The country code of the number it goes to. Meta publishes rates for 47 markets: 39 single countries and 8 regions, with an "Other" rate for the rest. A message to a number starting +91 is priced for India, one to +44 for the United Kingdom, whatever country your business is in.

So the same order update costs a different amount depending on where your customer lives, and a promotional message costs more than a plain update to the same person.

What is the 24-hour customer service window?

When a customer writes to you, a 24-hour window opens in which you can reply freely, without using an approved template.

This is the most important idea in WhatsApp pricing. WhatsApp treats a conversation the customer started very differently from one you started:

Inside the window, your team or AI assistant can answer in plain words: send a product link, a price, a photo, a delivery date. These are service messages. Every new message from the customer restarts the 24 hours.
Outside the window, anything you start must be a template: a message written in advance and approved by Meta. A restock alert the next day, an order confirmation, a payment reminder: all templates.

There is one more open door. When a customer messages you from a Click to WhatsApp ad (an ad on Facebook or Instagram with a "Send message" button), a free entry point window opens that can last up to 7 days. While it is open, Meta doesn't charge for your marketing, utility, authentication or service messages to that customer.

The practical lesson: wherever you can, let the customer start. A "Message us on WhatsApp" button on your website, a QR code at the counter, a number printed on the invoice: each one turns a paid message you would have started into a reply.

Inside the window you reply freely. Outside it, you need an approved template.

What are marketing, utility and authentication templates?

Templates are the messages you start, and Meta sorts each one into marketing, utility or authentication, each with its own rate.

Marketing templates promote something: an offer, a launch, a sale, a "we miss you" message. They usually have the highest rate, and they need the customer's clear agreement to hear from you.

Utility templates are about something the customer already did or asked for: an order confirmation, a shipping update, an appointment reminder, a payment receipt, a change to their account. They are usually cheaper than marketing.

Authentication templates carry one-time passwords for logging in or confirming an action.

The category is about what the message says, not what you call it. An order confirmation that ends with "and here's 20% off your next order" is a marketing message, charged as one. When you submit a template, Meta checks the category, and it can change it if the words don't match. Keeping updates purely about the order is the simplest way to keep them at the utility rate, and it's also what customers want from an update.

Which WhatsApp messages are free?

Messages customers send you are always free, and so are the first 1,000 service replies each month on every business number, and replies inside a Click to WhatsApp ad window.

Here is the full picture as of October 2026:

Messages from customers to you: never charged.
Service replies inside the 24-hour window: the first 1,000 delivered each month are free on every business phone number. Above that they are charged. The allowance resets every month and unused messages don't carry over.
Utility templates sent inside the window: free from July 2025, charged per message again since 1 October 2026.
Anything inside a free entry point window (after a Click to WhatsApp ad): free for up to 7 days.

One change matters for every business on the API: from 1 October 2026, Meta delivers service replies beyond the free 1,000 only if there is a payment method in Meta's Billing Hub. Without one, replies after the free allowance simply aren't delivered. If customers suddenly say they never got your answer, that is the first thing to check.

How do I estimate my monthly WhatsApp bill?

List every message you plan to send, put each in its category, count how many a month, and multiply by Meta's rate for your customers' country.

A worked way to do it, without guessing at prices:

List the moments. Order confirmed, shipped, out for delivery, delivered. Cash-on-delivery check. Appointment reminder. Login code. Monthly offer. Answers to questions.
Give each a category. Most moments in that list are utility. Login codes are authentication. The offer is marketing. Answers to questions inside the window are service.
Count them. Orders a month × updates per order. Cash-on-delivery orders × one check. Logins a month. Campaigns × the number of opted-in customers each one reaches. Service replies beyond the free 1,000.
Split by country. If most customers are in one country, one rate does most of the work. If you sell abroad, count each country separately.
Multiply by Meta's current rate card, linked under Sources, for each category and country.

Two refinements. Utility and authentication messages get volume tiers: past certain monthly volumes in a country, the rate per message drops, and the count starts again every month. And Meta bills Indian businesses in rupees since 1 January 2026; every messaging account of an eligible Indian business has to be moved to rupees by 31 December 2026, or its messages stop being delivered from 1 January 2027.

The number that usually surprises people is marketing. A single campaign to 20,000 opted-in customers is 20,000 marketing messages, more than many months of order updates. That isn't a reason to avoid campaigns; it's a reason to send them to people who asked for them and to measure what each one brings back.

Count each message by category, then multiply by Meta's rate for your customers' country.

How many messages can we send in a day?

A new business can start conversations with 250 different customers in any 24 hours, and the limit rises step by step as it proves it sends wanted messages.

This is Meta's messaging limit. It counts unique customers you start conversations with, not replies inside the customer service window. It applies to your whole business portfolio, so all your numbers share it.

The steps are 250, then 2,000, then 10,000, 100,000 and unlimited. You reach 2,000 by verifying your business with Meta (or having the partner who set you up verify it), or by delivering 2,000 good-quality messages to different customers within 30 days. After that, Meta raises the limit by itself, one step at a time, when your messages keep a high quality rating and you have used at least half of your current limit in the last 7 days.

Quality is the part you control. It drops when customers block you or report your messages, which mostly happens when people get messages they never asked for. Opt-in on record, easy opt-out and messages people actually want keep your quality high and your limit growing.

How do we keep WhatsApp costs under control?

Let customers start conversations, keep updates to the point, send marketing only to people who opted in, and read your numbers every month.

Invite the first message. Website buttons, QR codes and Click to WhatsApp ads turn messages you would pay for into replies.
Answer inside the window. An AI assistant or a shared inbox that answers within minutes keeps the conversation inside the 24 hours, so you don't need a template to follow up.
Write utility templates as utility. Order, booking and payment updates, with no offer tucked in at the end.
Choose which updates earn their place. A delivery update a customer would otherwise phone about pays for itself; a "thanks for visiting" message may not.
Target campaigns. Send offers to the customers likely to care, not the whole list, and stop sending to people who never open them.
Check the bill against results. Match messages sent to orders, bookings and fewer support calls each month, and drop what doesn't pay back.

Where have we built this?

We have connected the WhatsApp Business API to stores, clinics and delivery services, and built the shared inbox and AI agents a messaging platform's customers use.

For a skincare brand in Mumbai, every cash-on-delivery order is confirmed on WhatsApp before it ships, with the option to pay online in the same chat: one utility message that saves a wasted delivery. For the messaging platform we invest in, we built the shared team inbox and the flow and AI agent builders its customers run on WhatsApp. For a HealthTech startup we back in Ahmedabad, patients start the conversation themselves, booking video consultations on WhatsApp in Gujarati, so most of what follows are replies inside the window.

If you want to know what WhatsApp would cost and return for your business, the place to start is the list of moments in this guide. We can help you fill it in, and build the set-up once it adds up.

What decides your WhatsApp bill

There is no single price for the WhatsApp Business API: Meta publishes a rate card, and your bill follows what you send. These are the things that move it.

The category of each message. Marketing, utility and authentication templates each have their own rate. A message written as a plain update costs less than one that also sells.
Where your customers are. Rates follow the country code of the customer's number, so a business that writes to customers in several countries pays several rates.
Who starts the conversation. Replies inside the 24-hour window cost far less than messages you start, and replies to a Click to WhatsApp ad are free for up to 7 days.
How many you send. Utility and authentication messages get lower rates as your monthly volume in a country grows.
Your provider. Meta's charges are the same through any route; a provider or platform may add its own fee or monthly plan on top.
The build. Connecting WhatsApp to your store, CRM or ERP, the shared inbox, flows and an AI assistant are a one-time cost, separate from the messages.

Where we've built it

One utility message that saves a parcel (D2C Skincare Store): For a skincare brand in Mumbai, every cash-on-delivery order is confirmed on WhatsApp before it ships, and buyers can switch to paying online in the same chat.
Every conversation in one inbox (WhatsApp and SMS Messaging Platform): For the messaging platform we invest in, we built the shared team inbox its customers use to answer WhatsApp and SMS conversations from one place.
Customers who write first (Doctor Consultations in Gujarati): For a HealthTech startup we back in Ahmedabad, patients start the conversation by booking video consultations on WhatsApp, in Gujarati.

Questions and answers

Is the WhatsApp Business API free?

No, but parts of it cost nothing. Messages customers send you are never charged, the first 1,000 service replies a month per number are free, and replies to a customer who came from a Click to WhatsApp ad are free for up to 7 days. Messages your business starts are charged per delivered message.

Does Meta still charge per conversation?

No. Meta replaced conversation-based pricing with per-message pricing on 1 July 2025. Each delivered message has one category and is charged at most once.

Which category should an order update be?

Utility: a message about something the customer already did, such as an order, a booking or a payment. If the same message also promotes a product or an offer, Meta treats it as marketing and charges the marketing rate.

Do we pay for messages customers send us?

No. Meta charges only for messages your business sends, and only once they are delivered.

Can we pay Meta in rupees?

Yes. Meta started billing in Indian rupees on 1 January 2026 for businesses whose billing country is India, and asks them to move every messaging account to rupees by 31 December 2026. The currency is chosen when an account is created and can't be changed later.

How many customers can we message a day?

A new business can start conversations with 250 different customers in 24 hours. Verifying your business, or sending 2,000 good-quality messages within 30 days, lifts that to 2,000, and it then grows on its own to 10,000, 100,000 and unlimited while your messages stay high quality.

Is the WhatsApp Business app cheaper than the API?

The app is free, but it runs on a phone or two and can't connect to your software. The API costs per message and needs a build, and in return a whole team can share one number, updates go out by themselves and every chat is tied to the customer's orders.

Want one built for you? Sell and support on WhatsApp

Read the guide with its pictures and links on firelife.ventures: https://firelife.ventures/guides/whatsapp-business-api-pricing?utm_source=devto&utm_medium=republish&utm_campaign=guide-whatsapp-business-api-pricing

FIRELIFE Ventures builds software and AI agents for small and mid-size businesses, from Rajkot, India, for clients worldwide.

What an AI agent can do in a small business, and what it should never do alone

FIRELIFE Ventures — Sat, 14 Nov 2026 19:25:07 +0000

First published on firelife.ventures.

An AI agent can answer customers, read documents, update your software and chase follow-ups on its own, as long as a person stays in charge of money, judgement and anything unusual. The safest start is one frequent task with clear rules that you can easily check.

In short

An AI agent reads a request, checks your own systems, then answers, acts or hands the case to a person.
Good first jobs: order and booking questions, entering invoices and forms, follow-ups and first-line support.
Never alone: refunds and payments, prices and discounts, medical, legal or financial advice, and anything that can't be undone.
Start with one task that comes up often, follows clear rules and is easy to check, with a person approving answers at first.
It needs access to your tools, your rules in plain words and real past cases to be tested on.

What is an AI agent, in plain words?

An AI agent is software that reads a request, looks up what it needs in your own systems, and then answers, takes an action or hands the case to a person.

Think of a capable new assistant on their first week. They don't know your business by heart, but they know where to look: the order system, the booking calendar, the price list, the returns policy. When a customer asks something, they check before they answer. When something falls outside what they're allowed to do, they pass it to someone who can decide, with the details attached.

An agent works the same way, with three differences. It answers in seconds, at any hour. It never gets bored of the hundredth question about delivery times. And it keeps a written record of everything it read, did and decided, so you can always see why it answered the way it did.

What makes it an agent rather than just an AI that writes text is that it can do things: check an order, book a slot, update a record, send a reminder. What keeps it safe is that it only does the things you allow, with the access you give it.

How is an agent different from a chatbot or an automation?

A chatbot follows a script, an automation follows fixed steps, and an agent understands the request and decides which allowed step comes next.

A chatbot answers from a script or a list of questions. Ask it something slightly different and it either misunderstands or sends you in circles. Most people have met one and didn't enjoy it.

An automation (the kind you build with Zapier or inside your software) follows fixed steps: when an invoice arrives, copy these fields there. It's reliable when every case looks the same, and it breaks the moment one doesn't.

An agent sits between the two. It understands messy, real requests ("hi, I ordered the blue one last week, can I swap it for a medium?"), works out what's being asked, picks the right step from the ones you've allowed and checks the facts before it acts. When it isn't sure, it doesn't guess: it hands over.

In practice, good systems mix all three. Fixed steps where cases are always the same, an agent where they aren't, and a person where judgement is needed.

Which jobs can an agent take on in a small business?

The best jobs for an agent are the frequent ones that follow rules: answering routine questions, reading documents, updating records and following up.

Answering customers. Where is my order, can I change the delivery address, is the 4 pm slot free, do you have this in size L? An agent can answer these on your website, by email or through the WhatsApp Business API, from your real order and booking data, in the customer's own language. This is what most people mean by AI customer support.

Reading documents. Supplier invoices, purchase orders, forms, ID papers: the agent reads them, fills in the fields in your accounting or ERP software, checks totals against the order and flags anything that doesn't match. This is AI document processing, and it's often the job with the clearest saving of hours.

Updating your systems. Booking or moving an appointment, creating an order, logging a lead in your CRM, raising a support ticket. The agent does these through the same tools your team uses, with its own login and only the permissions it needs.

Following up. The quote nobody replied to, the invoice that's overdue, the customer who asked for a callback. An agent chases politely and on time, then passes the case to a person when it's time to talk.

Drafting for a person to send. Quotes, replies to tricky emails, summaries of a long thread. The agent prepares, a person reads and sends with one tap.

Answering your own team. A new hire asks how returns work. Instead of asking around, they ask an agent that knows your policies, and gets the steps with a link to the source.

A small shop, a clinic, a distributor and an agency would each pick different jobs from this list, but the pattern is the same: frequent, rule-based work moves to the agent, and judgement stays with people.

One message, five steps: the agent checks before it answers, and knows when to hand over.

What should an agent never do alone?

An agent should never decide alone on money, on judgement calls, on advice that needs a professional, or on anything that can't be undone.

These are the red lines we draw with every client, and we recommend you draw them too:

Refunds and payments. The agent can collect the details and prepare the refund. A person approves it.
Prices and discounts. It can quote your published prices. It shouldn't invent a deal, match a competitor or bend a rule because a customer pushed.
Medical, legal or financial advice. It can book the appointment, share your opening hours or send a form. The advice comes from a qualified person.
Anything that can't be undone. Deleting records, cancelling contracts, sending something to all your customers at once.
Anything unusual. An angry customer, a complaint about a staff member, a request that doesn't fit any rule. The agent's job here is to recognise the case and hand it over quickly, with everything the person needs.

This isn't a limitation to apologise for. It's what makes customers and staff trust the agent. A customer who gets a fast, correct answer about their order and a quick handover on their complaint is better served than by a person who takes hours for both.

What does an agent need from you?

An agent needs four things: access to the tools where your answers live, your rules in plain words, real past cases to learn and be tested on, and a person to hand over to.

Access to your tools. Your order system, booking calendar, accounting software, CRM, shared inbox: wherever the facts are. The agent gets its own account with the narrowest permissions that do the job, never a staff member's login.

Your rules, in plain words. What it may answer, what it may change, what it must hand over and to whom. Written like instructions for a new colleague, and agreed with you before it goes live.

Real past cases. A few hundred recent messages, documents or requests of the kind you want it to handle. They show how your customers actually write, and they become the test the agent must pass before it talks to anyone.

A person to hand over to. Someone who picks up the cases the agent passes on, ideally in the same tool your team already uses (WhatsApp, email, your helpdesk), not a new screen they have to remember to check.

You don't need perfect data or a technical team. Gaps show up quickly in testing, and the technical side is what the builder is for.

How do you pick the first task?

Pick the task that comes up most often, follows the clearest rules, is easiest to check and is cheapest to fix when something goes wrong.

Write down four or five jobs that take your team's time every week. Then ask four questions about each one:

Does it come up often? Many times a day is ideal. A job that happens twice a month isn't worth automating first.
Does it follow clear rules? If two of your staff would handle it the same way, an agent can learn it.
Is it easy to check? You should be able to look at the agent's answer and tell in seconds whether it's right.
Is a mistake cheap to fix? A wrong delivery estimate is an apology. A wrong refund is money.

The job with four ticks is your first task. For most small businesses it's order or booking questions, entering supplier invoices, or following up on enquiries. Refund decisions usually score one tick at most, which is exactly why they stay with a person.

Start there, measure it, and add the next task only once the first one runs smoothly.

Score each candidate on four questions. Start with the one that gets four ticks.

How do you keep it safe and under your control?

You keep an agent safe by limiting its access, writing its rules down, logging every step, testing it before launch and keeping a person in the loop until it has earned more freedom.

Least access. The agent can only see and change what its job needs. An agent that answers order questions doesn't need to see salaries.

Rules you can read. Not hidden in code: written down, agreed with you, and changed only with your say.

Every step on record. For each case, what it read, what it did and why. If a customer says the agent told them something odd, you can check in a minute.

Tested on your own cases. Before launch, it runs on real past cases and its answers are compared with what your team did. It goes live only when you're happy, and it's tested again after every change.

A person in the loop. At first, a person approves its answers before they go out. As it proves itself, you let it handle the simple cases alone and keep approval for the rest.

Your data stays yours. Use AI providers' business services, where your data isn't used to train their models, and keep data in the region you need.

How do you know it's working?

You know it's working when you measure the same few things every week: how many cases it handled without a person, how fast customers got an answer, how many answers were wrong, and how many handovers your team received.

Agree these measures before launch and note where you start from: how long replies take today, how many invoices are typed by hand, how many follow-ups slip. Then compare every week.

Watch the mistakes closely in the first weeks. Every wrong answer has a cause: a missing rule, a gap in your data, a question it hasn't seen. Fixing the cause, not just the answer, is what makes an agent get better over time.

And listen to your team. If the handovers arrive with everything they need, and the repetitive work has shrunk, the agent is doing its job, whatever the numbers say.

How do you start?

Start with one conversation about where your team's time goes, choose one task, test the agent on real cases with a person approving, and switch it on only when it passes.

List the repetitive work. What does your team do many times a day? Where do customers wait?
Choose the first task with the four questions above.
Collect the rules and past cases for that task.
Build and test the agent on those cases, connected to your tools with the least access it needs.
Run it beside your team with a person approving its answers, then let it handle the simple cases alone.

That is how our AI agent development work runs for small and mid-size businesses, from the first call to an agent that handles real work. If you'd like a second pair of eyes on which task to start with, write to us: we reply with our questions and a first plan.

What it costs, and what decides it

There is no fixed price for an AI agent: the cost follows the work you give it. These are the things that move it.

How many tasks it does. One clear task is the quickest and cheapest start. Every extra task adds rules, examples and testing.
Which tools it connects to. Software with a modern API connects quickly. Older or desktop-only software may need a small bridge first.
How much it handles. Running costs follow volume: the number of conversations, documents or tasks it handles each month.
Languages and channels. Each language and each channel (WhatsApp, website, email, phone) is set up and tested on its own.
How much a person checks. Having a person approve every answer at first costs less to build and earns trust. You widen what it does alone over time.

Where we've built it

AI agents on WhatsApp (WhatsApp and SMS Messaging Platform): For the messaging platform we invest in, we built the agent builder its customers use to answer order questions on WhatsApp and hand over to a person.
Visit notes the doctor approves (Pulse Clinic): In Pulse Clinic, our own clinic software, AI drafts the visit notes from the consultation, and the doctor checks and approves them.
Answers with the working shown (E-commerce Analytics Dashboards): For a home goods brand in Denver, an agent answers questions about sales and stock in plain English and shows the query behind every figure.

Questions and answers

Can a small business afford an AI agent?

Yes, if it starts small. One clear task on tools you already use is the cheapest way in, and running costs follow how much the agent handles. You see it work before adding the next task.

Do we need clean data before we start?

No. The agent needs access to where the answers already live: your order system, bookings, price list or policies. Gaps and messy spots show up when it is tested on real past cases, and they are fixed then.

Should customers know they are talking to an AI agent?

Yes. We recommend saying so in the first message and letting customers ask for a person at any time. It builds trust, and some countries require it.

What happens when the agent doesn't know the answer?

It says so and hands the case to a person, with the conversation and what it already checked. An agent that guesses is a badly built agent.

Is an AI agent the same as ChatGPT?

No. ChatGPT is a general assistant you talk to. An agent is built around your business: it runs on a model like the ones behind ChatGPT or Claude, but answers from your own data, works inside your tools and follows your rules.

Do we need a developer to run it?

Not day to day. Your team reads the handovers and approves drafts. Changes to its rules and connections are technical work, done by whoever built it or by your own developer.

Which task should a small business hand to an agent first?

One that comes up many times a day, follows clear rules, is easy to check and is cheap to fix if it goes wrong. For most businesses that is order or booking questions, invoice entry or follow-ups.

Want one built for you? Put an AI agent to work in your business

Read the guide with its pictures and links on firelife.ventures: https://firelife.ventures/guides/ai-agents-for-small-business?utm_source=devto&utm_medium=republish&utm_campaign=guide-ai-agents-for-small-business

FIRELIFE Ventures builds software and AI agents for small and mid-size businesses, from Rajkot, India, for clients worldwide.