Free Service Messages on WhatsApp End October 1, 2026: The H2 2026 Meta Pricing Overhaul

Meta just published upcoming pricing rules for non-template messages — the ones you send in reply to a user inside the 24-hour customer service window. The rules take effect between August and October 2026, and they materially change the WhatsApp cost structure for every business that does customer support, AI replies, or reactive messaging on the platform.

Here is what is changing, in order of when it hits your bill.

1 August 2026 — Meta Business Agent Platform starts billing

Meta launches the Meta Business Agent Platform on 1 July 2026 and starts charging for it on 1 August 2026. Any partner or directly integrated client can integrate. The platform is Meta's own AI response engine — you connect it via API, users message you on WhatsApp, Meta's AI generates the reply.

The important nuance: Meta charges once for Meta Business Agent messages, covering both the AI compute and the message delivery. Businesses using a third-party AI (OpenAI, Anthropic, etc.) integrated via a BSP will pay twice — once to the AI provider, once to Meta for delivery.

1 October 2026 — Service messages become paid

Since 1 November 2024, service messages (non-template replies sent by a human agent or third-party AI inside an open customer service window) have been free. That ends on 1 October 2026. Every service message will then be charged at the same per-message rate as utility and authentication messages for that market.

For India, that means every reply your support team sends inside a CSW starts costing roughly ₹0.115 per message from Oct 2026. At scale, that adds up fast. A support team handling 100,000 replies/month goes from ₹0 to ~₹13,000/month just for message delivery — before any AI or platform costs.

Meta explicitly notes there are no volume tiers for service messages, unlike utility and authentication. You do not unlock lower rates at higher volume.

1 October 2026 — Utility templates inside the CSW become paid

The second half of the 1 October change: utility templates delivered inside an open customer service window — free since 1 July 2025 — become paid at the standard utility rate. For India, that is ~₹0.115 per message.

Concretely: today, if a customer messages you asking about their order and you reply with an order-confirmation utility template within 24 hours, that template is free. From 1 Oct 2026, it costs ₹0.115.

The one free window that survives — Free Entry Point (FEP)

The 72-hour Free Entry Point window remains free for message delivery. When a user contacts you via a Click-to-WhatsApp Meta Ad or a Facebook Page CTA button (on Android/iOS), you get a 24-hour customer service window plus a 72-hour FEP window from the moment you reply. Inside the FEP window, all messages you send — template or non-template — are free of delivery charges.

The exception: if the reply is generated by Meta Business Agent, you still pay for token usage even inside the FEP window. Only message delivery is free in FEP, not AI compute.

How the math shifts for a real Indian SMB

Consider an Indian D2C store handling ~20,000 customer support replies per month (about 700/day) inside customer service windows, powered today by a human team.

For high-volume, low-complexity support (order status, "where is my package"), service messages via human/BSP-AI are dramatically cheaper. For lower-volume, complex multi-turn interactions (troubleshooting, product recommendations), Meta Business Agent's per-token pricing can undercut a third-party-AI-plus-BSP-delivery combo.

What Indian businesses should do this quarter

1. Front-load anything that relies on free CSW utilities. If your Diwali / Q4 order-confirmation flow depends on free utility templates inside the customer service window, that free ride ends 30 September 2026. Book those campaigns for August-September.
2. Audit your Click-to-WhatsApp Meta Ads funnels. The 72-hour FEP window is now the only free delivery window that survives. Every reply, template or not, is free inside it. If you are not building your campaigns around FEP windows, you are leaving real savings on the table.
3. Model your Oct 2026 support budget — take your current monthly support-reply volume and multiply by the service-message rate for your recipient country (~₹0.115 for India, higher for many other markets). Plus 18% GST. That is your new floor cost from Oct 2026, before AI or platform fees.
4. Reconsider "AI everywhere" strategies. Handing every customer reply to a third-party AI now costs twice — AI provider fee + Meta delivery. Reserve AI for interactions that genuinely need it; use structured template flows where possible, since templates have volume tiers and AI-generated replies do not.
5. Decide on Meta Business Agent vs third-party AI. Meta Business Agent's one-charge model is genuinely simpler (and often cheaper for complex interactions) but locks you into Meta's AI. Third-party AI (OpenAI, Anthropic, custom models) gives flexibility but doubles the invoice.

How this connects to the other WhatsApp pricing changes

The H2 2026 overhaul stacks on top of two other India-specific deadlines you should already be tracking:

• INR billing migration by 31 Dec 2026 for India-based partners — or messages stop delivering from 1 Jan 2027.
• The 1 January 2026 marketing template rate hike for India (already in effect).

Combined, an Indian business that has not planned for any of this could see their real WhatsApp cost line go up 30-80% in 2027 versus 2025 — silent deliverability failure on top of new per-message charges on top of higher marketing rates.

Avianya AI is executing the INR migrations on a rolling schedule through Q3-Q4 2026, absorbing the volume-tier webhook setup for customers, and building out modelling for Oct-2026 service-message costs so you can see your projected new bill before it hits. Book a 15-minute call if you want us to run the numbers for your current volume.