# AI chatbot pricing models explained URL: https://instantknowledgebot.com/learn/ai-chatbot-pricing-models Last updated: 2026-10-14 > AI chatbot pricing usually follows one of five models: flat plans with message allowances, credit systems that vary by AI model, per-conversation fees, per-resolution (outcome-based) fees, or bring-your-own-key, where you pay the AI provider directly. The cheapest option depends on your message volume, how predictable it is, and which features you need. **AI chatbot pricing** can be confusing because vendors measure usage in different units: messages, credits, conversations, resolutions or AI tokens. Two products with similar headline prices can cost very different amounts for the same traffic. This explainer describes the common pricing models, their trade-offs, and how to estimate what you'll actually pay. ## What are the main AI chatbot pricing models? | Model | You pay for | Predictability | Typical fit | |---|---|---|---| | **Flat plan with message allowance** | A monthly fee that includes N AI messages | High | Small and mid-sized websites | | **Credits** | Credits per month; each AI reply costs credits, often more for advanced models | Medium | Teams that want model choice | | **Per conversation** | Each conversation in which the AI replies | Medium | Helpdesk and live chat platforms | | **Per resolution (outcome-based)** | Each issue the AI resolves, by the vendor's definition | Low to medium | Enterprise customer service | | **Bring your own key (BYOK)** | A platform fee, plus your AI provider's token charges | Medium | High volume, model choice, developers | Many products combine models, for example a plan with included credits plus paid top-ups. ## How do message allowances work? You pay a fixed monthly price and get a number of AI-answered messages. Watch for: - **What counts as a message:** usually one visitor question answered. - **What happens at the limit:** the bot stops, switches to a fallback, or overage charges apply. - **Top-ups:** whether extra messages expire. **Pros:** easy to budget. **Cons:** you may pay for unused capacity, or hit the cap in a busy month. ## How do credit systems work? Credits are a flexible currency. One reply might cost one credit on a basic model and several credits on an advanced one. That lets vendors offer many models under one plan, but makes it harder to know how many messages you'll get. Read the credit table for the model you'll use, and check whether credits roll over. ## How does per-conversation pricing work? You pay for each conversation in which the AI replies, regardless of how many messages it contains. Tidio's Lyro, for example, prices AI conversations this way, with a free allowance to start (as listed on Tidio's AI agent page, checked 2026-10-14). It's intuitive for support teams, who think in conversations, but long conversations and short ones cost the same. ## How does per-resolution pricing work? You pay when the AI resolves an issue, as defined by the vendor, often when the customer doesn't ask for a human afterwards. Enterprise platforms such as Intercom (Fin) and Zendesk describe outcome-based pricing for their AI agents on their own sites. It aligns cost with value, but: - **Definitions vary** between vendors, so read them closely. - **Costs scale with success:** the better the bot, the higher the bill. - **Budgeting is harder** when you don't know resolution rates in advance. ## How does bring-your-own-key pricing work? You connect your own API key from an AI provider such as OpenAI, Anthropic, Google or xAI. The platform charges for its software; the provider charges for AI usage by tokens. Benefits: no markup on AI usage, model choice and, on some platforms, no message cap. Downsides: two bills and variable AI costs. Full explainer: [bring your own key (BYOK) explained](/learn/bring-your-own-key-explained). ## How do you estimate your real chatbot cost? 1. **Estimate monthly questions.** Look at your inbox, chat logs or site traffic. For a new bot, start with a conservative guess and adjust after a month. 2. **Translate into the vendor's unit:** messages, credits (at your chosen model), conversations or resolutions. 3. **Add required features.** Lead capture, branding removal, team seats, analytics and extra bots are often tied to higher plans. 4. **Add add-ons:** extra message packs, extra bots, white-labeling. 5. **Consider the yearly discount** if you're confident in the tool. 6. **Compare the cost per useful answer**, not just the plan price. ## What hidden costs should you look for? - **Branding removal** sold as a separate add-on. - **Extra bots or agents** charged individually. - **Seats** for teammates. - **Model multipliers** in credit systems. - **Overage rates** above your allowance. - **Setup or onboarding fees.** ## How does InstantKnowledgeBot price its plans? We use a simple allowance model plus bring-your-own-key: - **Free plan:** 1 bot and 100 hosted AI messages a month, with unlimited messages on your own AI key. - **Paid plans (Starter, Pro, Agency):** more bots, larger monthly hosted-message allowances and features such as lead capture, CSV export, branding removal, team seats and client workspaces. Billed monthly or yearly. - **Bring your own key on every plan:** Claude, ChatGPT, Gemini or Grok, with no message cap from us and no markup. - **Message packs:** one-time purchases of extra hosted messages that never expire. - **Free trial:** 14 days of Pro, no card, starting when you create your first bot. What counts: one visitor question answered by our hosted AI is one message, and allowances reset on the 1st of each month (UTC). When the allowance runs out, bots use message packs, then your own key if connected. See current prices on the [pricing page](/pricing), set up a key with [our BYOK guide](/guides/bring-your-own-ai-key), or compare vendors in our [roundup of AI chatbot builders](/compare/best-ai-chatbot-builders). ## Which pricing model is right for you? - **Low, steady volume:** a flat plan or a free tier. - **High or spiky volume:** BYOK or a plan with non-expiring top-ups. - **Enterprise support desk:** per-resolution or per-conversation pricing inside a helpdesk may fit. - **Agencies with many clients:** plans with many bots and client workspaces, with BYOK for heavy clients. Whatever you choose, run a pilot for a month and measure usage before committing to a yearly plan. Compare chatbots with people too: [AI chatbot vs live chat](/learn/ai-chatbot-vs-live-chat). ## Frequently asked questions **How much does an AI chatbot cost per month?** It ranges from free plans for low volume to hundreds or thousands of dollars a month for enterprise support platforms. The pricing model, your message volume and the features you need matter more than headline prices. **What is a message credit?** A unit some platforms use to meter AI usage. One message may cost one credit on a basic model and several credits on a more advanced one, so the number of messages you get depends on the model. **What is per-resolution pricing?** An outcome-based model where you pay when the AI resolves a customer's issue, as defined by the vendor, rather than per message or per seat. It's common in enterprise customer service platforms. **Is bring-your-own-key cheaper?** Often at higher volumes, because you pay provider rates with no markup and no platform cap. At low volume, a plan's included messages can be simpler. **How does InstantKnowledgeBot price its plans?** A free plan, plus Starter, Pro and Agency plans with monthly hosted-message allowances, bot limits and features, billed monthly or yearly. Bring-your-own-key is available on every plan with no message cap. Current prices are on the pricing page.