WhatsApp is About to Start Charging for Your Replies: What the October 2026 Service Message Change Costs You

A shopper messaged, “Is this back in stock?” Your agent replied; a short back-and-forth followed, and the whole exchange sat comfortably within a free 24-hour window. That era is ending.

From July 2025 to September 2026, both service messages and utility templates sent inside the 24-hour window were free for businesses using the WhatsApp Business Platform. Before July 2025, utility templates were charged even inside the window. Non-template service messages were free from roughly November 2024 to September 2026.

On 1 October 2026, Meta reintroduces charges for service messages: the everyday, free-text replies your team sends inside that 24-hour window. It is part of the biggest reshuffle of WhatsApp Business pricing since Meta moved from conversation-based to per-message billing (a shift Wati broke down in its guide to conversation-based pricing).

If your business runs support, sales, or marketing over the WhatsApp API, the unit economics of every chat are about to change, and the businesses that understand the details now will pay a lot less than the ones that find out in October.

Here is exactly what is changing, why Meta is doing it, what it will cost, and the practical moves that keep your bill in check.

Wait – What Is Actually Changing?

Today, when a customer messages you first, you get a 24-hour customer service window in which non-template free-form replies (service messages) are free. Utility templates sent within that window were also free — but only from July 2025 to September 2026; from 1 October 2026, they will be charged again.

According to Meta’s official developer documentation, as of 1 October 2026, service messages become billable, and utility templates lose their in-window waiver. (Authentication templates were never free inside the window — they’ve been charged all along.)

Here is how the free-and-charged history actually breaks down by message type:

Meta is splitting non-template replies into two categories with two different price tags:

Service messages – replies sent by a human agent or a third-party AI tool. Billed per message from 1 October 2026.

Meta Business Agent messages – replies generated by Meta’s own AI agent. Billed per token from 1 August 2026.
Two important reassurances up front.

• This affects only the WhatsApp Business Platform (API), including clients using it via business partners (BSPs) and On-Premises API implementations. It does not affect the consumer WhatsApp app or the WhatsApp Business app used without the API.

• Conversations that open from a Click-to-WhatsApp ad or a Facebook CTA button still get their free 72-hour entry-point window.

One nuance worth knowing: that free window covers message delivery, if you use Meta’s own AI to reply inside it, the token charge still applies. The launch was reported by India Today.

First, What Counts as a “Service Message”?

This is the part most teams get wrong, so it is worth nailing down. A service message is a reply that is all three of these things at once:

It is free-form (not an approved template), sent within the open 24-hour window, and not generated by the Meta Business Agent.

If you send any of these as plain text from a human agent or a non-Meta AI tool, they are service messages, and from October, each one is billable:

• “Your order has shipped and should arrive on Thursday.”
• “Can you share a screenshot of the error you’re seeing?”
• “An agent will join this chat in about two minutes.”
• “To verify your account, please share the last 4 digits of your number.”

These are the operational, human replies that make up the bulk of real support conversations. Until 30 September 2026, they are free. After that, they are line items.

The Three Dates That Matter

Sources: Meta developer documentation; India Today; Business Today (July 2026).

How Much Will It Actually Cost?

Here is the rule to remember: a service message will cost the same per-message rate that Meta already charges for utility and authentication template messages in that country. So the price you pay depends on where your customers are.

Meta will publish the exact October rates by 1 September 2026, but today’s template rates give you a reliable ballpark:

India: utility and authentication messages sit around ₹0.115 per message today; service messages are expected to match, per Business Today.

Brazil (Meta’s own example): message delivery is about 0.68 cents (roughly $0.0068) per message.

Volume tiers: unlike templates, service messages get no volume discount – the per-message rate stays flat no matter how many you send.

Rule of thumb: if you know your country’s utility template rate, you already know roughly what a service message will cost from October. Multiply that by the number of free-text replies your team sends each month, and you have your new line item.

Meta Business Agent vs. Service Messages: Two Very Different Meters

The change that trips people up is that Meta is now metering AI and humans differently.

Meta Business Agent (Meta’s AI) is billed per token from 1 August 2026 at a single global rate of $2.00 per million tokens. A typical exchange runs about 20,000–25,000 tokens – roughly 4–5 cents a message (per Meta’s documentation; Business Today notes a message can use up to 25,000 tokens). Simple replies cost less; complex ones cost more.

Meta’s own Brazil illustration makes the gap concrete:

Meta’s illustrative Brazil example (per 10,000 messages); actual costs vary by country and message volume. Source: Meta developer documentation. Confirm current figures before budgeting.

Why Is Meta Doing This?

Three reasons, and they are not hard to read.

1.The 24-hour service window has been free for nearly two years, an enormous volume of business traffic that Meta has not been charging for.

2.Reintroducing service-message fees monetizes that window directly. Second, the token model lets Meta price AI by the actual compute a conversation consumes, rather than a flat per-message fee.

3.Per-message pricing nudges businesses to send fewer, better messages — the same anti-spam logic behind Meta’s earlier 2024 rate restructuring, which TechRadar covered at the time. Fewer, higher-quality messages protect the experience that keeps people on WhatsApp in the first place.

What This Quietly Does to Your Support Costs

The headline rate looks tiny. The math is where it bites. Because you are now charged per reply, a chatty conversation becomes more expensive.

A worked example: A customer asks about a delivery. Your agent sends 6 free-text replies to resolve it. Today: free. From October 6, service-message charges for that one conversation. Now multiply across 5,000 support chats a month, each averaging 5–6 replies — and a line item that used to read “$0” becomes very real.

How to Keep Your WhatsApp Costs Under Control

You do not have to absorb this blindly. The businesses that come out ahead will make a handful of practical changes before October:

Resolve in fewer turns. A single, well-structured reply beats five fragmented ones. Consolidating multiple messages into one directly reduces what you pay.

Move data collection into WhatsApp Flows. Capturing a name, address, or preference in one interactive step replaces several charged back-and-forth messages.

Use templates where they fit. Order updates, OTPs, and status alerts belong in utility or authentication templates – predictable, compliant, and often the smarter unit cost.

Protect your free windows. Route more first-touch demand through Click-to-WhatsApp ads, which keep the free 72-hour entry-point window.

Let AI handle the repetitive 80%. An AI agent that answers common questions end-to-end keeps humans (and their per-message replies) focused on the conversations that actually need them.

If you want the full picture on what the API costs before and after this change, Wati’s WhatsApp pricing guide and its breakdown of what the WhatsApp API really costs in 2026 are good companions to this article.

How Wati Helps You Get Ahead of the Change

This is the moment a WhatsApp platform stops being a nice-to-have and starts paying for itself.

Wati is built to help marketing, sales, and support teams do more with fewer, smarter messages – which is exactly what the new pricing rewards.

Astra, Wati’s AI agent, resolves routine questions end-to-end across WhatsApp, web, and voice -qualifying leads, booking meetings, and closing tickets in a turn or two instead of a long human thread. See how it works on the Astra product page.

No-code chatbots and WhatsApp Flows collect information in one interactive step, so you are not paying for a chain of follow-up replies.

Broadcasts, templates, and analytics in one place help you shift structured updates to templates and see exactly where your message spend goes.

Cost visibility and controls -see exactly where your message spend goes across categories, so you can model the October change and manage it before it hits your bill.

Curious which features matter most this year? Wati’s roundup of the top WhatsApp features for brands in 2026 is a useful next read.

Get Ahead of October Before It Hits Your Bill

See how Wati’s AI agent, WhatsApp Flows, and unified inbox help you resolve more in fewer messages – and keep your WhatsApp costs predictable as the new pricing lands.

Book a demo at wati.io

This guide is based on publicly available information as of July 2026. Meta publishes final per-country rates by 1 September 2026 – always confirm current rates in Meta’s official documentation and with your BSP before budgeting.

Frequently Asked Questions