Until October 2026, anyone running a WhatsApp chatbot in India could rely on one comforting fact: replies inside the 24-hour customer service window were free. Agents chatted, chatbots answered, support tickets closed, and nothing got added to the Meta invoice except the odd outbound template.
That is no longer true. Since 1 October 2026, service messages have a free tier of 1,000 per business phone number per calendar month, and every reply beyond that is billed at ₹0.115. For a modest WhatsApp chatbot handling even 500 customer conversations a month with 4–5 messages each, that's thousands of billed service messages the business was not paying for in September.
This guide explains exactly what counts as a service message, how the 1,000 free tier works, four worked examples across business sizes, and the levers that actually lower your monthly chatbot bill.
Quick answer
A WhatsApp chatbot's monthly bill comes from three line items: service messages, templates sent outside the 24-hour window, and any Marketing or Utility templates it fires. As of November 2026 the rates are:
| Message type | Rate (INR) | Notes |
|---|---|---|
| Service first 1,000 per month | Free | Per business phone number, resets on day 1 |
| Service after 1,000 | ₹0.1150 | Every free-form reply inside the 24-hour window |
| Utility template | ₹0.1150 | Billed even inside the customer service window (new rule) |
| Marketing template | ₹0.8631 | Any promotional template the bot fires |
| Authentication template | ₹0.1150 | OTPs sent through the bot |
A chatbot serving 2,000 conversations/month with 4 replies each sends ~8,000 service messages. Subtract 1,000 free = 7,000 billed × ₹0.115 = ₹805 per month in service messages alone, plus whatever templates the bot sends. Always confirm current rates on Meta's official WhatsApp Business Platform pricing page.
Platform policy: WhatsApp requires opt-in before business messaging, honouring opt-outs, and approved templates for Platform-initiated conversations outside the 24-hour customer service window. Read Meta's current Business Messaging Policy.
Contents
- What is a service message?
- How the 1,000 free tier actually works
- What your chatbot costs: 4 worked scenarios
- Hidden costs most chatbot projects miss
- Chatbot vs human agent: which costs more per month?
- How to lower your chatbot bill without hurting the experience
- Frequently asked questions
- Conclusion
What is a service message?
A service message is any free-form (non-template) message your business sends inside the 24-hour customer service window a user has opened by messaging you first. Anything you send in those 24 hours that is not a pre-approved template is a service message.
Meta counts the following as service messages:
- Automated chatbot replies to inbound queries
- AI-generated replies from an LLM-backed bot
- Agent replies typed in a shared team inbox
- Menu-based bot flows that answer FAQs
- Any attachment, image, or document sent free-form inside the window
Meta does NOT count the following as service messages:
- Inbound customer messages (you never pay for these)
- Messages sent outside the 24-hour window — those need templates and are billed by template category
- Replies sent inside the 72-hour free entry-point window opened by Click-to-WhatsApp ads or Facebook Page CTAs — those are free regardless of volume
The practical implication: if your bot nurtures cold leads via Click-to-WhatsApp ads, almost all of that traffic is free for the first 72 hours per conversation. If your bot answers support queries from customers who find your number on your website or packaging, every reply after the first 1,000 of the month is billed. Check the current platform documentation before you build.
Related reading: WhatsApp Business API: What It Is, How It Works & Setup
How the 1,000 free tier actually works
The mechanics matter, because this is where most teams get confused:
- Per business phone number, not per WABA, not per BSP account. Each phone number connected to WhatsApp Business Platform has its own counter.
- Resets on the 1st of every calendar month at midnight UTC. Not on your billing cycle. Not 30 days from signup.
- Counts outbound service messages only. Inbound customer messages are not counted.
- Counts after delivery. If a message fails to deliver, it is not counted.
- Not transferable. Unused free messages do not roll over to next month.
- Same rate for everyone after 1,000. No volume tiers on service messages.
A real pattern: a business runs two numbers — one for sales, one for support. The sales number sends 200 service messages a month (well under the free tier). The support number sends 3,500 service messages a month (2,500 billed). The free tier does not pool: the support number pays for its own overflow and the sales number's unused 800 free messages disappear at month-end.
Teams that split traffic across many numbers often end up paying more than teams on a single, well-managed number with a shared team inbox.
What your chatbot costs: 4 worked scenarios
Scenario 1: A small D2C brand, support-only chatbot
500 customer queries/month, avg 5 bot replies each, 50 agent escalations with 4 replies each. Single phone number.
- Bot replies: 500 × 5 = 2,500 service messages
- Agent replies: 50 × 4 = 200 service messages
- Total service: 2,700. Minus 1,000 free = 1,700 billed × ₹0.115 = ₹195.50
Essentially free. The bot pays for itself if it saves a single agent hour.
Scenario 2: A growing coaching institute, lead-qualification chatbot
3,000 inquiries/month via Click-to-WhatsApp ads (free entry-point, 72h free), 1,500 inquiries from website widget (not free), bot sends avg 6 replies each.
- Click-to-WhatsApp replies: 3,000 × 6 = 18,000 messages, all free for 72 hours
- Website widget replies: 1,500 × 6 = 9,000 service messages. Minus 1,000 free = 8,000 billed × ₹0.115 = ₹920
- Utility confirmations for booked demo classes: 400 × ₹0.115 = ₹46
- Total: ₹966
Click-to-WhatsApp ads carry most of the volume for free. Moving the website widget traffic to Click-to-WhatsApp ads where possible would cut the bill further. Pair with booking automation for the demo-class slot.
Scenario 3: An e-commerce support chatbot at scale
15,000 customer queries/month across one support number, avg 8 bot/agent replies each.
- Service messages: 15,000 × 8 = 120,000. Minus 1,000 free = 119,000 billed × ₹0.115 = ₹13,685
- Utility templates for shipping updates (sent proactively, not inside the window): 20,000 × ₹0.115 = ₹2,300
- Total: ₹15,985
The chatbot still beats the cost of 4–5 full-time support agents by 10×, but the service-message line is now meaningful. Volume here justifies investing in a smarter bot that resolves tickets in fewer turns.
Scenario 4: A fintech app with OTP + support bot
500,000 OTPs/month, 10,000 support queries with avg 4 bot replies each.
- Authentication templates: 500,000 × ₹0.115 = ₹57,500
- Service messages: 10,000 × 4 = 40,000. Minus 1,000 free = 39,000 billed × ₹0.115 = ₹4,485
- Total: ₹61,985
At 750,000+ Authentication messages/month, volume tier discounts kick in and OTP cost drops. The support-bot line stays small relative to OTP spend.
Related reading: Build a WhatsApp AI Agent That Drives Sales
Hidden costs most chatbot projects miss
- BSP platform fees. A chatbot runs on a platform. That platform usually charges a monthly subscription (₹2,000–₹25,000+) or a per-conversation fee on top of Meta's ₹0.115.
- AI token costs. If your bot uses GPT, Claude, or any LLM behind the scenes, you pay API tokens per message. A chatbot processing 10,000 turns a month on GPT-4-class models can add ₹5,000–₹30,000 in token fees.
- Human agent handoff cost. When the bot escalates to a human, that agent's time is a real cost. Factor in salary + overhead per ticket.
- Template re-approvals. If your bot sends utility or marketing templates and Meta reclassifies them, you lose runway and pay the new (usually higher) rate.
- Failed-escalation templates. If a user drops off and the bot sends a 24-hour-later re-engagement, that template is billed as Marketing (₹0.8631) if it has any promotional wording.
- Multi-number sprawl. Teams that spin up new numbers for every campaign fragment their 1,000 free tier. Consolidate before scaling.
- Chatbot analytics tools. Third-party analytics for conversation quality and funnel drop-off often sit outside the BSP and have their own subscriptions.
Chatbot vs human agent: which costs more per month?
A rough back-of-the-envelope comparison for an Indian team handling 5,000 customer queries/month on WhatsApp:
| Setup | Monthly cost (INR) | Notes |
|---|---|---|
| Chatbot only (5 replies/query) | ₹2,760 (service) + BSP fee + AI tokens | Lowest cost, lowest flexibility |
| Chatbot + 1 agent for escalations | Chatbot cost + ~₹25,000 agent salary + overhead | Best balance for most SMBs |
| 3 human agents only | ~₹90,000 salary + ₹2,760 service messages + overhead | Highest cost, most flexibility |
Message fees are rarely the biggest line. Agent salaries dwarf service-message costs at any Indian business. The right question is not "how do I avoid the ₹0.115 per message" but "how do I resolve more tickets in fewer turns with the right mix of bot + human".
Related reading: How to Measure WhatsApp Marketing Results: Key Metrics
How to lower your chatbot bill without hurting the experience
- Design conversations to resolve in fewer turns. A bot that answers in 3 replies instead of 7 cuts service-message cost by more than half.
- Consolidate support to one phone number. Each number has its own 1,000 free tier; you lose the overflow across multi-number setups.
- Route cold acquisition through Click-to-WhatsApp ads. 72-hour free entry-point means the bot nurtures for free.
- Batch answers in a single message. Instead of three sequential replies ("Sure!" → "Let me check" → "Here is your answer"), send one complete reply.
- Use smart segments to only route high-value queries to the bot. Low-intent leads can get a template reply outside the window.
- Add an FAQ step before the bot engages. Static FAQ buttons resolve 30–40% of queries without entering the paid loop.
- Audit template categorisation quarterly. A utility template that got reclassified to Marketing without you noticing can 7.5× your template cost overnight.
- Trigger follow-ups inside the 24-hour window, not with a new Marketing template 48 hours later.
- Time your broadcasts to open a customer service window first, so any downstream nurture is service messages, not Marketing templates.
- Review your invoice line-by-line monthly. Billing mistakes happen — compare Meta's line items to your BSP's.
Frequently asked questions
Quick answers related to this guide.
What is a service message on WhatsApp?
A service message is any free-form (non-template) message your business sends inside the 24-hour customer service window a user opened by messaging you first. This includes chatbot replies, agent replies, and AI-generated replies.
How many free service messages do I get on WhatsApp?
1,000 per business phone number per calendar month, resetting on the 1st. After that, each service message costs ₹0.115 in India.
Do chatbot replies count as service messages?
Yes. Meta does not distinguish between a human agent typing a reply, a rule-based chatbot responding, or an AI-powered bot generating an answer. All are billed as service messages.
What does a WhatsApp chatbot cost per month in India?
Meta's service-message cost alone ranges from ₹0 for a small bot handling 500 queries/month to ₹10,000+ for a large support bot handling 15,000+ queries. Add BSP platform fees and any AI/LLM token costs on top.
Are Click-to-WhatsApp ad chatbot replies free?
Yes, for 72 hours after the user first messages you through the ad. Any reply (bot or agent) in that window is a free entry-point message and does not count against your 1,000 free service-message tier.
Does the 1,000 free tier roll over?
No. Unused free service messages do not carry over to the next calendar month.
Can I pool the free tier across multiple numbers?
No. Each business phone number has its own independent 1,000 free tier. Consolidating traffic to one number usually saves money.
Which is cheaper chatbot or human support agents?
Chatbot messaging fees are always cheaper than agent salaries at any Indian business scale. The honest trade-off is experience quality and ticket complexity, not cost.
