AI chatbot for your website: what it costs and how setup works

“How much does an AI chatbot cost?” has a frustrating answer: anywhere from free to six figures. The spread is real because “chatbot” covers everything from a canned FAQ widget to a fully integrated assistant that looks up orders and books meetings. This guide breaks the options into clear cost bands, shows what actually drives the price, and walks through how a real setup happens.

What you’re actually paying for

A website chatbot has three cost layers, and vendors blur them:

• The interface — the chat widget itself. This is the cheap, commoditized part.
• The intelligence — how well it understands questions and pulls the right answer from your content. This is where quality and cost diverge.
• The integrations — connections to your CRM, help desk, order system, or knowledge base. This is usually the biggest driver of both price and value.

A chatbot that only answers from a fixed FAQ is cheap because it skips layer three. One that can tell a customer where their order is has to connect to real systems — and that’s what you’re really paying for.

Cost by option (2026)

Watch the pricing model, not just the sticker. Per-seat tools (common with help-desk suites) multiply fast — a 10-agent team can run $850/month before AI fees are added. Per-resolution pricing is honest about outcomes but gets expensive at high volume. Custom integrations typically add $5,000–$25,000 each, and integration work can add 20–50% to a build budget.

What actually drives the price

Two chatbots described the same way can differ 10× in cost. Here’s what moves the number:

• Integrations. Reading from or writing to your CRM, order system, or help desk is the single biggest driver — each connection adds build and testing time.
• Knowledge quality. Clean, well-organized help content is cheap to train on. Scattered, outdated, or contradictory content needs a cleanup phase first.
• Accuracy bar. “Good enough” deflection is affordable. The last few points of reliability — where the bot handles edge cases without a wrong answer — are where cost climbs.
• Actions vs answers. A bot that answers is a chatbot. A bot that acts — checks an order, updates a record, books a slot — is closer to an AI agent, and costs accordingly.
• Compliance. Regulated data (health, finance, PII) needs extra controls, logging, and human-in-the-loop review.

How setup actually works

Whether you buy or build, a good setup follows the same arc:

1. Pick the job. Define the top questions or tasks the bot should handle — usually the highest-volume, most repetitive support requests. Scope beats ambition here.
2. Gather the knowledge. Point the bot at your help center, docs, and past tickets. Clean and de-duplicate first; the bot is only as accurate as its source.
3. Connect the systems it needs — CRM, order lookup, calendar. This is where timelines stretch, so start early.
4. Set the handoff. Decide exactly when the bot escalates to a human, and make that path fast. A bot that traps people is worse than no bot.
5. Test on real questions from your actual inbox, not invented ones. Measure a baseline before launch.
6. Launch narrow, then widen. Start with one topic or one page, watch the transcripts, and expand once it earns trust.

A platform bot on existing content can go live in days to two weeks. A custom build or one with deep integrations typically takes 4–8 weeks including testing.

Platform vs custom: how to decide

The math usually favors starting on a platform. Against a $79–$300/month SaaS tool, the break-even on a $30,000 custom build can be years — unless the platform genuinely can’t do what you need. Build custom when you have unusual integrations, strict data-control requirements, or behavior no off-the-shelf tool offers. Otherwise, buy to validate, and revisit custom only if you hit a real wall. This is the same logic we cover in build vs buy: custom AI agent vs off-the-shelf tools.

One more distinction worth making before you spend: if your “chatbot” really needs to complete tasks — not just answer — you’re shopping for an agent, and the cost and setup change. See AI agent vs chatbot and what is an AI agent. If it’s one step inside a larger flow, our guide to Zapier vs Make vs AI agents shows how the pieces fit together.

Related guides

• AI agent vs chatbot — which do you actually need?
• Zapier vs Make vs AI agents: which automation tool fits
• What is an AI agent? A plain-English guide
• Our services: AI Chatbots · AI Agents · Web Development

Sources & further reading

• IBM — What is a chatbot?
• Zendesk — AI chatbots for customer service
• Intercom — pricing (per-seat / resolution reference)

Frequently asked questions