Chatbot pricing is confusing on purpose. Five vendors will quote you five numbers that cannot be compared, because each one meters something different: one counts messages, one counts conversations, one counts successful outcomes, one counts unique people, and one counts seats. The sticker prices look similar. The bills do not.
This is a decoder. It explains each metering model, shows what each one does to your bill when your traffic doubles, and gives you a method for turning any vendor's pricing page into a single comparable monthly number.
The five models
| Model | You pay for | Bill spikes when | Predictability |
|---|---|---|---|
| Message credits | Each message or AI action | Chats get longer, or actions fire per turn | Poor |
| Conversations | Each unique thread in a time window | Traffic rises, even from browsers who never buy | Medium |
| Resolutions / outcomes | Each problem the AI solves | Your bot gets better | Poor |
| Contacts (MAU) | Each unique person per month | A post goes viral | Poor |
| Flat plan with a stated cap | A monthly fee, with a published ceiling | Nothing, until you cross the cap | High |
1. Message credits
You buy a pot of credits and each interaction draws it down. The trap is that a "credit" is rarely one visitor message. Depending on the vendor, one visitor turn can consume several credits when the bot calls a tool, looks something up, or runs a multi-step action. Chatbase auto-recharges at roughly $40 per 1,000 credits when you run out, which is the part that turns a $120 plan into a $200 month.
How to read it: ask the vendor, in writing, how many credits a typical five-turn conversation with one lookup consumes. If they cannot answer plainly, assume three to five times what you expected.
2. Billable conversations
You pay per unique thread within a window, usually 24 hours. Tidio works this way, and separately meters its AI agent as an add-on, so an AI-handled chat can tick both meters. Fairer than credits, because a chatty visitor and a terse one cost the same. Still exposed to raw traffic: someone who opens the widget, reads one line and leaves can be billable.
How to read it: your billable conversations are roughly your monthly widget opens, not your leads. Multiply your site traffic by 2 to 5% to estimate.
3. Per resolution or outcome
Intercom's Fin charges $0.99 per resolution on top of seat fees ($29 to $132 per seat per month). The pitch is compelling: you only pay when it works. The problem is structural. The better your bot performs, the higher your bill, so success is a cost centre and there is no ceiling. A two-seat Advanced team with 500 resolutions a month is $170 in seats plus $495 in resolutions, about $665 a month.
How to read it: forecast monthly resolutions and multiply by the per-resolution price, then add seats. Under 200 resolutions it is affordable. Over 1,000 it has left small-business territory. Also ask who decides what counts as "resolved", because that definition is the price.
4. Per contact (monthly active users)
ManyChat's model. You pay by unique contacts reached in a month. Reasonable for a stable audience, brutal for anyone doing social media, because a single reel that performs can push you two tiers up overnight for traffic that included no buyers at all.
How to read it: look at your worst month of the last year, not your average, and price that.
5. Flat plan with a stated cap
A monthly fee, a published message allowance, and no surprise meters behind it. PopABot works this way: Starter free with 300 messages a month, Essential $19/month with 1,000, Pro $49/month with 8,000, Enterprise 30,000. When you approach the cap you upgrade deliberately rather than discovering it on an invoice.
How to read it: the only question is whether the cap fits. Divide the allowance by the average messages per conversation (usually 5 to 8) to get conversations per month.
The four costs that never appear in the comparison table
Sticker price is the beginning of the number, not the end. Four line items do more damage to a chatbot budget than the plan fee itself:
- Branding removal. Getting the vendor's badge off your widget is a paid upgrade almost everywhere. Chatbase sells it as a separate $1,188/year add-on. Tidio requires the Plus tier, which starts around $749/month. PopABot includes it on every paid plan from $19/month. For a small business this single line often exceeds the entire plan cost elsewhere.
- Extra chatbots or agents. Most platforms give you one bot and charge for the second. Chatbase charges roughly $300/year per additional agent. If you run two brands, a client site, or a separate bot for support and sales, multiply accordingly. PopABot bundles 1 / 2 / 5 / 999 by tier.
- Seats. Per-seat platforms bill every person who touches the inbox. A three-person team on a $29 seat price is $87/month before the AI does anything.
- The integrations you assumed were included. SMS notifications usually mean a Twilio account plus Zapier. Calendar booking usually means Calendly. Each is $10 to $30 a month and a fragile chain of automations. Check whether the features you are actually buying the tool for are native or bolted on.
A method for comparing any two vendors honestly
Do this on paper before you talk to anyone's sales team. It takes fifteen minutes and it is the only way to get comparable numbers.
- Estimate your monthly conversations. Site visitors × 2 to 5% is a reasonable widget-open rate for a small business site. Be honest and use a busy month.
- Multiply by 6 to get approximate messages, if the vendor meters messages rather than conversations.
- Write down your non-negotiables. Branding removed? SMS to your phone? Calendar booking? More than one bot? Each becomes a line item, priced at that vendor.
- Price the busy month, not the average. Any usage-metered plan should be evaluated at your peak, because that is the month the invoice hurts.
- Add year-one total, not monthly. Annual add-ons like branding removal distort monthly maths badly. Total cost of year one is the comparable figure.
- Ask what happens at the cap. Three possible answers: the bot stops, it keeps going and you are billed, or you are prompted to upgrade. All three are acceptable; not knowing which is not.
A worked example
A trades business with 4,000 monthly website visitors, wanting branding removed, SMS alerts and one chatbot. Assume 3% widget-open rate, so about 120 conversations, roughly 720 messages a month.
| Platform | Plan needed | Branding removal | SMS alerts | Approx. year 1 |
|---|---|---|---|---|
| PopABot Essential | $19/mo (1,000 messages) | Included | Included (10/mo) | $228 |
| Chatbase Standard | $120/mo | +$1,188/yr | Twilio + Zapier | $2,628+ |
| Tidio + Lyro | ~$24/mo + $32.50/mo add-on | Plus tier only | Not native | $678+, badge still visible |
| Intercom Advanced, 2 seats | $170/mo seats | Included | Native | $2,040 + $0.99 per resolution |
Pricing verified as of May 2026; vendors change pricing frequently, so re-check before you buy. The point of the table is not the exact figures, it is the shape: the gap between platforms is driven by add-ons and metering, not by the headline plan price. Full breakdowns in the seven-platform comparison, the Chatbase alternatives guide and the Tidio alternatives guide.
Six questions to ask before you sign anything
- What exactly consumes one unit of my allowance, in a five-turn conversation with one lookup?
- What happens when I hit the cap: stop, overage, or upgrade prompt?
- Is removing your branding included on this plan, and if not, what does it cost per year?
- How many chatbots does this plan include, and what does the next one cost?
- Which of my required features are native and which need Zapier or a third-party account?
- If I cancel, what happens to my training content and my conversation history?