To give you full transparency on your operating costs, this article breaks down the billing model applied to AI interactions.
The billing model: per answer
Unlike regular messages, which can be billed per unit based on Meta sessions, the AI Agent runs on a cost per generated answer.
Single rate: Each answer delivered by the AI Agent is billed 0,05 € (5 cents).
The "grouped answer" rule
WhatsApp sometimes enforces technical limits on message length or display to keep messages readable on mobile. For this reason, the AI Agent may split a single answer into two or three consecutive messages.
The key point to remember:
No matter how many text bubbles (messages) are sent to answer one specific customer question, you are only billed once.
Concrete example:
A customer asks a complex question about your delivery times.
The AI replies with:
Message 1 : "Hello! Our delivery times depend on your location."
Message 2 : "For France, allow 48 hours. For Europe, allow 4 to 5 business days."
Total billed for this interaction: 0,05 €.
Why this model?
This per-answer billing system has several advantages for your store:
Predictability: You know exactly what helping a customer costs you, no matter how wordy the AI gets.
WhatsApp optimization: The AI is free to structure its messages for maximum readability without increasing your costs.
Cost efficiency: At 5 cents per answer, handling a support ticket costs drastically less than a human agent, while delivering an instant reply.
Tracking your usage
You can track AI Agent spending directly in your Kanal interface:
Go to Settings > Message Billing.
The total amount consumed by the AI is included in your overall balance and charged through your usual Auto-refill system.
Note on Meta fees: The AI Agent billing (0,05 €) covers Kanal's AI service.