AI Chatbot Pricing in 2026: Per-Resolution vs Per-Conversation (Intercom, Tidio, Chatbase Compared)
What does an AI support chatbot actually cost? We pulled the live pricing pages for Intercom Fin, Tidio Lyro, and Chatbase in July 2026 and did the per-conversation math. Per resolution, per credit or per conversation with a limit you set — the billing model matters more than any feature difference.
AI chatbot pricing is confusing by design. One vendor charges per "resolution," another per "conversation," a third per "message credit" — and none of those words mean the same thing.
So instead of comparing marketing pages, we pulled the live pricing pages of three popular AI support chatbots in July 2026, re-checked them in October 2026, and did the arithmetic. Every number below links to the source, so you can verify it yourself. The short version:
Monthly cost of 500 AI-handled conversations
Competitor pricing checked October 2026 · one website · bars to scale
+ seat costs · $0.99/outcome
at the $0.78/convo entry-tier rate, monthly billing
Standard plan credits, monthly billing
500 × $0.10 per AI conversation + Starter $5, under a limit you set
Same job. 9× price spread before Intercom's seat costs. The difference isn't features — it's the billing model.
The three billing models
Per-resolution. You pay every time the AI successfully answers someone. Intercom Fin works this way. The better the bot performs, the more you pay.
Per-credit / per-conversation quotas. Your plan includes a bucket of AI conversations or message credits; overage costs extra. Tidio Lyro and Chatbase work this way.
Per-conversation, with a limit you set. One price per conversation the AI answers, nothing when it hands off before answering, and a cap the owner sets for each billing period. GhostChat works this way. A busy month can’t surprise you.
What each vendor actually charges (re-checked October 2026)
Intercom Fin — $0.99 per outcome, plus seats
Intercom's pricing page lists Fin at $0.99 per outcome — charged when a customer confirms resolution, stops replying, or Fin completes a workflow. That's on top of per-seat plan pricing, and add-ons like Copilot run $29/agent/month.
The math scales against you: 500 Fin outcomes a month is $495 in Fin fees alone — before a single seat is paid for. Per-resolution pricing means your bill grows precisely because the product is working.
Tidio Lyro — from $39/mo for 50 AI conversations
Tidio's pricing page sells the Lyro AI agent as a separate quota: paid Lyro plans start at $39/mo for 50 AI conversations with monthly billing ($32.50/mo billed annually). That's 78 cents per AI conversation at the entry tier — on top of the base Tidio plan (Starter is $29/mo with 100 billable live conversations).
You can technically add Tidio to multiple sites, but per their own FAQ, the widget shares one appearance and behavior config across all of them — there's no per-site AI knowledge base on a single sub.
Chatbase — message credits, and $1,188/yr to remove branding
Chatbase's pricing runs on message credits: Hobby is $40/mo for 700 credits, Standard is $150/mo for 4,000 (both monthly billing; yearly billing is $32 and $120), and auto-recharge overage costs $40 per 1,000 credits. A single AI reply can consume multiple credits depending on the model.
The add-ons are where it compounds: removing the Chatbase branding from your widget is $1,188 per year, and each extra AI agent in your workspace is $25/month.
GhostChat — $0.10 per conversation, on any paid plan
GhostChat's pricing doesn't sell the AI as a plan. Switch the assistant on from Starter ($5/mo) or Team ($25/mo) and it costs $0.10 per conversation. You pay $0.10 once the AI gives a real answer. If it hands off before answering anything, it's free. A chat counts once even if you take it over later, and a visitor back after 24 hours of silence starts a new one. You set a spending limit for each billing period ($20 by default; raise it, lower it or switch it off). You set the limit. We never charge a cent over it. Lowering it doesn't refund chats the AI already answered. At the limit the chat keeps working and hands off to you. No credit packs, and no branding-removal add-on on Team. Free covers live chat without the AI.
The per-conversation math
Here's what a month of 500 AI-handled conversations on one site costs on each product, using the verified numbers above:
| Product | Billing model | ~Cost for 500 AI convos/mo | Effective per convo |
|---|---|---|---|
| Intercom Fin | $0.99/outcome + seats | ~$495 + seat costs | $0.99 |
| Tidio + Lyro | Base plan + AI quota | ~$29 base + Lyro tier sized for 500 convos (entry tier is 50 convos for $39) | ~$0.78 at entry tier |
| Chatbase | Message credits | $40–$150 depending on credits per reply | $0.04+/credit, multiple credits per reply |
| GhostChat | Per conversation, limit you set | $55 on Starter (500 × $0.10 + $5/mo) | $0.10, free when it hands off first |
At 500 conversations a month, the spread between the cheapest and most expensive option is 9× before Intercom's seat costs. No feature gap in this market is that large.
The multi-site multiplier
Metered pricing gets dramatically worse if you run more than one site — an agency with client sites, or a founder with a few products. Per-resolution and per-credit models multiply with every property, and most vendors scope a subscription (or at least the AI's knowledge base) to a single site or workspace.
On GhostChat, one account covers unlimited sites — Team is $25/mo for 10 seats, each site gets its own widget config, welcome message and AI FAQ knowledge base, and the AI is $0.10 per conversation across all of them under one spending limit. The same footprint on per-site or per-resolution pricing runs into hundreds or thousands per month.
Where the expensive options earn it (honest caveats)
Intercom is a full enterprise suite. If you need Salesforce workflows, SLAs, and a large support org, Fin’s price rides on genuinely deep tooling. Small sites are paying for infrastructure they’ll never touch.
Meters exist everywhere — including ours. GhostChat charges per conversation too, at $0.10. The difference: you set the monthly limit, hand-offs before an answer are free, and at the limit visitors still reach you live or by email — never a surprise bill.
AI doesn’t replace you. Every product on this list hands off to a human when the bot is out of depth. Pricing should reflect the bot’s share of the work — which is exactly why paying per successful answer feels backwards.
Why a dime a conversation is possible
Modern small language models made AI support cheap to serve. GhostChat's assistant runs on Claude Haiku with aggressive prompt caching — a typical grounded answer costs us well under a cent. Vendors charging $0.78–$0.99 per AI conversation aren't covering compute; they're pricing against the salary of the human you'd otherwise hire.
That's a fine business model. It's just not one you're obligated to fund — especially if you're a small site where the "salary saved" framing never made sense to begin with. See how the assistant works on our AI chatbot page, or compare plans on pricing.
AI support without the surprise bill
$0.10 per conversation the AI answers, on Starter or Team. Nothing else to buy, no credit packs, and a spending limit you set.
More from the blog
AI Chatbot vs Live Chat: Which Does Your Website Actually Need?
The "bot or human" choice is a false one for most small sites. An honest breakdown of what AI chatbots are genuinely good at, where live chat wins, and why you probably want both — without paying enterprise prices for it.
The Best Privacy-First Tools for Your Website in 2026
Replace Google Analytics, Intercom, and Typeform with tools that don't track your visitors. Analytics, chat, forms, email, and hosting — all privacy-first.
We Re-Tested 5 Chat Widgets in 2026 — Opening Intercom Downloads 833KB
We re-ran our February benchmark with a rebuilt harness: median-of-5 runs, three scenarios, and a metric nobody publishes — what it costs when a visitor actually clicks the chat bubble. Intercom ends up at 1.1MB. Four other widgets downloaded hundreds of KB and rendered no chat at all.