You bought an AI chatbot to answer your customers. Six months in, it got good. It handles the sizing questions, the shipping questions, the "does this come in blue" questions. And your bill went up. If that sounds familiar, the problem is not your chatbot. It is the pricing model underneath it.
There are four common ways store-chat vendors charge you. Knowing which one you are on tells you what your bill will do as the bot improves and as your traffic grows.
The four pricing models, in plain terms
- Flat monthly tier. You pay a fixed amount each month for a bucket of chats or features. Predictable until you cross the next tier, then it jumps. You often pay for headroom you do not use.
- Per-seat. You pay per human agent who has a login. Fine for a support desk full of staff. For an online store where the whole point is that the bot does the talking, you are paying for seats you may not need.
- Per-resolution. You pay every time the bot resolves a question on its own, often $0.90 to $1.00 each. Sold as "you only pay for outcomes." Read the next section before you sign.
- Per-conversation. You pay one flat rate for a whole chat, however many questions the shopper asks inside it.
Why per-resolution punishes a good bot
Per-resolution sounds fair. You pay when the bot does its job. But look at what "its job" means over time.
A new chatbot resolves maybe half of what shoppers ask. The rest gets handed to you or your team. As you feed it better content and it learns your catalog, it starts resolving more. That is the whole reason you bought it. Under per-resolution pricing, every extra answer it handles is a new charge. The better it gets, the more it costs. You are billed for your own success.
Now add a busy season. Black Friday, a viral product, a holiday rush. Your chat volume can double or triple in a week. Under per-resolution, your bill doubles or triples with it, in the exact month your margins are tightest and your time is shortest. There is no natural ceiling unless you go and cap it by hand, and many tools make that hard to find.
There is one more quiet cost. Some helpdesk tools count a resolution as a ticket too, so the same interaction shows up in two places on your invoice. You end up paying twice for one shopper.
What per-conversation pricing does instead
Per-conversation charges one flat rate for the whole chat. A shopper can ask about sizing, then shipping, then returns, then add a product to their cart, all in one conversation, and it counts once. The bot getting smarter does not move your bill. It just resolves more inside the same flat charge.
That flips the incentive. You can tune the bot to answer more, hand off less, and walk more shoppers to checkout, without watching a meter climb. A good bot and a basic bot cost the same per chat. The only thing that changes your spend is how many real conversations you have, which is the one number you want to pay for.
How Emporiqa prices it
Emporiqa is pure pay-as-you-go. $0 per month to start, $0.25 per conversation, and the AI model costs are included in that price, so there is no separate bill for usage and no key to manage. New stores get $25 of signup credit, about 100 conversations, and no card is asked for at signup.
The worst case is capped on purpose. Every account has a $59 per month default cap that you can raise or lower yourself from the billing dashboard. So even in your busiest month, you decide the ceiling before it happens. No surprise invoice, no per-seat fee, no annual lock-in. The bot answering one question or ten inside a chat costs you the same $0.25.
A chatbot that gets better at selling should make you more money, not cost you more to run. Per-conversation pricing keeps those two things on the same side.
Want to see the numbers on your own store? Create a free Emporiqa account with $25 of signup credit, about 100 conversations, and no card required. Or watch it work first on the live demo at demo.emporiqa.com.
More on the model that removes this trap: a pay-as-you-go AI chatbot that bills per conversation.
