Most content about AI chatbot pricing is written by chatbot vendors. The numbers are either vague ("plans start from $X") or structured to make the paid option look good. I build custom chatbots for businesses, so I have no incentive to push you toward any particular option. Here is a real breakdown.
The three actual options
No-code SaaS chatbot (Tidio, Intercom, Freshchat, etc.): $30 to $300 per month depending on features and conversation volume. You configure it through a visual interface, no code required. The bot works, handles FAQ-style questions well, and can be live in a day. The downside is that it is generic. It does not know anything specific about your business unless you train it manually, and even then the training is limited. It also costs more as your conversation volume grows.
No-code AI builder (Voiceflow, Botpress, etc.): $50 to $500 per month. More customisable than out-of-the-box SaaS. You can build more complex conversation flows and connect to your own data. Still no code required if you stay within the platform's constraints. Gets expensive fast if you have volume, and you are still somewhat limited by what the platform allows.
Custom chatbot built on an LLM API: Build cost of $1,500 to $8,000 depending on complexity, then ongoing API costs of roughly $5 to $50 per month depending on usage. You own the code, can connect it to anything, and can tune the behavior precisely. No platform markup. The tradeoff is upfront build cost and the need for a developer when you want to change something substantial.
What the custom option actually costs to run
The OliviaGRE chatbot runs on Claude Haiku via the Anthropic API, served through a Netlify Function. At around 500 conversations per month (each roughly 5 to 8 turns), the API cost works out to about $3 to $8 per month. Netlify's free tier covers the serverless function at that volume. Hosting cost is effectively zero.
That is the honest running cost for a small business chatbot on a custom build: a few dollars a month. The cost that matters is the build, not the ongoing API spend.
What determines build cost
The bot itself (system prompt, basic response logic, a widget on the page) is maybe 2 to 3 days of work. What takes longer is everything around it:
- Logging conversations to a database or sheet so you can see what people are asking
- Capturing email inside the conversation (versus sending people to a form)
- Connecting the bot to downstream systems: booking tools, CRM, email platforms
- Rate limiting and abuse protection so the API does not get hammered
- Handling edge cases where the bot should escalate to a human or a different page
That full integration adds another 1 to 3 weeks depending on how many systems are involved. The OliviaGRE chatbot connects to MailerLite for email capture and has logic to route between lead magnet delivery and trial lesson booking. That took longer than the bot itself.
Which option is right for your situation
| Situation | Best option |
|---|---|
| Need something live this week, FAQ-style answers, low volume | No-code SaaS, around $30 to $50/month |
| Want custom conversation flows, have some technical patience | No-code AI builder like Voiceflow |
| Want the bot to connect to your actual data, CRM, bookings | Custom build, $2,000 to $5,000 upfront |
| High conversation volume where per-conversation pricing hurts | Custom build, API costs scale cheaply |
| Just want to answer "what are your prices" and not much else | No-code SaaS, or honestly just a good FAQ page |
The question worth asking before spending anything: what specific job should the chatbot do? "Add a chatbot to the website" is not a brief. "Reduce DMs asking about pricing by capturing them in chat and routing to the booking page" is a brief. A vague chatbot gets ignored. A chatbot with a clear job pays for itself.
Frequently asked questions
Want a chatbot that actually converts?
I build custom AI chatbots connected to your booking system, CRM, and email tools. Free call to see if it makes sense for your business.
Book a Free Call →