How much does an AI chatbot cost?

AI chatbot pricing looks confusing because vendors charge in very different ways. Once you understand the handful of pricing models, comparing what you will actually pay — today and as you grow — gets a lot easier.

The common pricing models

• Free / freemium: fine for a quick trial, but usually capped on messages, training content or branding, and rarely enough for real support.
• Flat monthly subscription: a fixed price per month for a set capacity. Predictable — your bill does not jump because the bot had a busy week.
• Per-resolution or per-message: you pay each time the bot answers. This scales your cost directly with usage, so the more successful the bot is, the more you pay.
• Per-seat: priced by agent, common for tools bolted onto a help desk.

The catch with usage-based pricing

Per-resolution pricing sounds fair — pay for what works — but it has an awkward incentive: your bill grows in lockstep with the volume you wanted to deflect. A bot that resolves 10,000 questions a month is doing its job, but on per-resolution pricing that success is also your biggest line item.

What to actually compare

1. Your expected volume: estimate monthly questions, then model the cost under each pricing model at that volume — and at 2× and 5×.
2. What is included: training content limits, number of agents, human handoff, integrations, data residency.
3. The all-in cost: add setup time and overage fees, not just the headline price.

Predictable beats cheap-looking

For most teams the steadiest choice is flat monthly pricing: you know the number, and the savings from deflection stay with you as volume grows. That is how Chatixy prices — flat monthly plans with no per-resolution fees — so the agent gets *more* valuable as it does more, not more expensive. See the current plans on the pricing page.