Quick answer: The AI messaging agent uses metered charges with per message cost applying. You're charged when your AI agent responds to an incoming text message, and the cost depends on how long the message exchange is and which AI model you've selected.
Understanding AI Messaging Agent Pricing
DialLink's AI messaging agent charges are based on the AI responses your business sends, not on the messages you receive. Think of it like this: every time a customer texts your business and the AI generates a reply, that response gets charged based on a few factors.
You only pay for what you use. If you receive 10 customer texts in a month and the AI responds to all of them, you're charged for those 10 responses. If you receive 100 texts and AI responds to all of them, the cost scales accordingly.
What Affects Your Costs
Message and Conversation Length
Longer messages cost more than shorter ones. A brief "Thanks! We'll call you back soon" costs less than a detailed response with multiple paragraphs of information.
The cost is based on the total length of the conversation visible to the AI, including:
AI agent’s prompt length
The conversation history
The length of the AI agent's reply
If a customer sends multiple texts and the conversation continues, each new AI response includes the full conversation history. Longer conversations naturally cost more per response than single-message interactions.
Which AI Model You Choose
DialLink supports different AI models, and each has its own pricing tier. Some models are more powerful (and cost more), while others are faster and more affordable. When you set up AI messaging agents, you'll choose the model that works best for your business.
You're not locked into one choice. You can switch models anytime in your settings. This gives you the flexibility to find the right balance between capability and cost for your needs.
Language and Text Type
If your customers text primarily in English using standard characters, the cost is straightforward. However, if you support multiple languages or use non-Latin scripts (like Arabic, Chinese, or Cyrillic), processing that text costs more. This means conversations in these languages may cost a bit more per message than English conversations.
It's not a major difference, but it's worth knowing if you serve multilingual customers or operate in non-English-speaking markets.
What Triggers a Charge
A charge happens when all three are true:
A customer sends your business an incoming text (SMS, WhatsApp message, or MMS with a text caption)
Your AI messaging agent is enabled for that phone number
The AI messaging agent generates and sends a response
If the AI doesn't respond (for example, because you've disabled AI messaging agent or the message doesn't match your AI agent's settings), there's no charge.
Each charge is tied to the reply sent for a single incoming message. Even if the same customer sends the same message twice, each one triggers a response and a separate charge.
Example Calculation of AI Messaging Agent Usage Charges
The billing unit is one SMS segment; however, when setting up the AI agent, the pricing for each model is presented per 1,000 SMS segments.
Now let’s look at the example calculation. The purpose of this is to demonstrate the logic of how the charges are calculated. Your real charges will differ from this example, as they depend on multiple factors described above.
The AI messaging agent replies mid-conversation, using an approximately 850-character system prompt, which counts as about 6 segments. The conversation history consists of 10 messages of approximately 100 characters each, or around 10 segments. The incoming message being replied to counts as 1 segment, and the reply is around 300 characters, or about 2 segments. Altogether, this totals around 19 segments. The table below shows the total cost of this reply, by model.
Model | Cost of that one reply |
SONNET_4_5 | $0.0814 |
GPT_5 / GPT_5_1 | $0.0543 |
HAIKU_4_5 | $0.0271 |
GPT_5_MINI | $0.0109 |
GPT_5_NANO | $0.0022 |
What Other Charges Are Associated with Messaging
Domestic and international inbound and outbound messages carry their own charges. Because AI messaging agents send and receive messages, these charges apply in addition to your AI messaging agent charges. For details, see our support article on metered charges.
Common Questions
Will I be charged if my team also replies?
No. Charges only apply when the AI generates a response. If your team manually replies to a customer text, that doesn't incur an AI messaging agent charge. It's just a normal business message and is charged accordingly.
How do I keep costs predictable?
Because pricing is based on usage, costs naturally stay lower if conversations are shorter or less frequent. A few ways to think about this:
Keep your AI agent instructions concise. Shorter instructions mean lower cost per response.
Monitor conversation patterns. If you notice long back-and-forth exchanges, you might adjust your AI's instructions to resolve issues faster, in one or two messages instead of many.
Start small and scale. If you're new to AI messaging agents, you'll get a natural sense of your usage and costs in the first few weeks. There's no setup fee or minimum. You only pay for what you use.
What if I'm worried about unexpected costs?
That's a fair question. The best way to get comfortable is to start with a limited rollout:
Enable AI messaging agents on one phone number for after-hours replies first
Monitor your usage for a week or two
Check your billing to see what you're actually spending
Then decide if you want to expand to more numbers or enable it more broadly
Your account shows detailed usage breakdowns, so you can always see exactly what you're being charged for.
Can I disable an AI messaging agent to stop charges?
Yes, absolutely. You can turn AI messaging agents on or off anytime. Once it's off, no new charges accrue.