Chatbot build costs in 2026: $3k to $15k for a rule-based bot, $15k to $50k for an LLM support bot, and $60k+ for generative AI. Timeline, running costs, and budgeting.
A custom chatbot costs between $3,000 and $500,000 in 2026, and the number depends on what the bot does, not on any single price list. A rule-based FAQ bot runs $3,000 to $15,000, an LLM support bot grounded in your own knowledge base runs $15,000 to $50,000, and a generative AI assistant that answers from your documents starts around $60,000 and can pass $200,000. Most mid-market projects land in the $15,000 to $50,000 band and ship in 6 to 12 weeks. Treat these as planning ranges, not quotes.
| Chatbot type | Typical build cost | Timeline | What you get |
|---|---|---|---|
| Rule-based FAQ bot | $3,000 – $15,000 | 3 – 6 weeks | Decision-tree answers and lead capture |
| LLM support bot (RAG) | $15,000 – $50,000 | 6 – 12 weeks | Answers from your docs, helpdesk integration |
| Generative AI assistant | $60,000 – $200,000+ | 4 – 8 months | Multi-source Q&A, actions, analytics |
| Enterprise AI agent | $150,000 – $500,000+ | 6 – 12 months | Multi-system autonomous workflows |
Five things move a chatbot quote more than anything else. The first is integration depth. A bot that only answers from a static FAQ page is cheap because there is nothing to connect it to. A bot that reads your CRM, creates support tickets, and checks order status needs real engineering, and that is where an $8,000 bot becomes a $35,000 bot.
The second is data readiness. Most of the work in a modern chatbot is not training a model. It is cleaning and structuring your documents, support transcripts, and product data so the model can retrieve the right information. Companies with scattered or messy documentation spend more time here than anywhere else.
The third is the model and where it runs. A hosted API such as OpenAI or Anthropic Claude keeps upfront cost low but adds per-token usage fees every month. Self-hosting an open model cuts usage fees but raises infrastructure and engineering cost. The fourth is channels: a web widget only is simple, while WhatsApp, Slack, and voice each add integration and testing. The fifth is compliance. A chatbot handling payments or health data triggers security and review work that a marketing FAQ bot never sees.
A rule-based bot follows a fixed decision tree. It is fast to build and predictable, but it breaks the moment a customer phrases a question differently. These suit FAQ pages, appointment booking, and lead capture.
An LLM chatbot reads intent and generates answers from your own content, usually through retrieval-augmented generation (RAG). It costs more because someone has to prepare the knowledge base, wire up the retrieval pipeline, and tune the prompts and guardrails. This is where the $15,000 to $50,000 mid-market range comes from. If you want the bot to also act, such as cancel an order or update a record, the price climbs again because each action is an integration with a system you already run.
The build price is only part of it. LLM API usage, hosting, and monitoring typically run $150 to $3,000 a month for a production bot, depending on conversation volume and which model you use. A busy support bot answering thousands of conversations sits at the higher end. Add someone to review transcripts and update the knowledge base when your product changes, and the total cost of ownership over a year often matches or exceeds the original build. When you compare vendors, ask for the monthly figure, not just the one-time quote.
We start from the buyer's question, not the technology. A chatbot built to deflect support tickets is a different project from one that closes sales or runs an internal HR helpdesk, and the scope follows the job. We usually begin with a two- to four-week discovery phase to inventory your documents, define the channels, and pick between a hosted LLM and a self-hosted model based on your data rules. Then we build a small RAG pipeline first so you can test the answers on real questions before we expand the integrations. Our team works in Java, Python, and Node, and has built these systems for clients across several markets. If you already have a support or sales workflow in mind, our AI development services cover the full build, and this chatbot build guide walks through what a working bot needs.
Start with one job and one channel. A bot that answers your ten most common questions on your website is an $8,000 to $15,000 project you can measure in weeks, and you do not need an agent that routes across five systems on day one. Use a hosted model first; you can self-host later if volume makes API fees expensive. Prepare your data before you brief a vendor, because a clean FAQ and a tidy knowledge base cut the discovery phase and the build. Finally, buy when the problem is generic and build when the answer depends on your own data. A SaaS chatbot tool at $50 to $400 a month handles simple FAQ deflection, while a custom build earns its cost when the bot must understand your specific documents and connect to your systems. For a fuller view of what custom AI costs, see our AI development cost guide.
A rule-based FAQ bot costs $3,000 to $15,000 to build and takes three to six weeks. It answers fixed questions from a decision tree and is the cheapest option when the goal is basic deflection and lead capture.
An LLM chatbot grounded in your own data typically costs $15,000 to $50,000 for a support bot with one or two integrations. A generative AI assistant with RAG over multiple systems starts around $60,000 and can pass $200,000.
Plan on three to six weeks for a rule-based bot, six to twelve weeks for an LLM support bot with integrations, and four to eight months for a generative AI assistant. Data preparation and integrations, not coding, drive most of the timeline.
Buying a SaaS chatbot at $50 to $400 a month is cheaper when the bot answers generic questions. Building is worth it when answers depend on your own documents, or when the bot must act inside your CRM, helpdesk, or ordering system.
LLM API usage, hosting, and monitoring usually run $150 to $3,000 a month for a production bot, plus staff time to review answers and update the knowledge base. Over a year these costs often match the original build price.
Yes. Zendesk, Intercom, HubSpot, Salesforce, and similar systems expose APIs that a custom bot can read from and write to, so it can create tickets, look up orders, and update records. Each integration adds cost, which is why most teams start with one or two.
Planning a chatbot for support, sales, or internal use? Talk to us about your project and we will help you scope the features, channels, data preparation, and a realistic budget.
GlobeSoft builds custom software, AI applications, SaaS, FinTech and enterprise systems for clients worldwide. Message us on WhatsApp for the fastest answer, or email us — we typically reply within 24 hours.
Keep reading
A realistic breakdown of ERP implementation costs in 2026: licensing, implementation services, data migration, and the hidden costs vendors don't quote.
8 min readA Flutter app costs $15,000 to $45,000 for a simple build and up to $250,000+ for a complex one in 2026. What moves the price, and how to keep it down.
7 min readAn LMS costs $3 to $25 per user per month to license, or $50,000 to $300,000+ to build custom. Here's how to budget for either path.
9 min readA plain-English breakdown of monthly dedicated team pricing by region, team size, and seniority, plus what pushes the number up and how to keep it down.
5 min readAI chatbot development usually fails on scope, not on the model. Why narrow bots win, where they pay off, and how to build one customers actually use.
7 min readWhat it costs to build an e-invoicing system in 2026, the formats and clearance models that drive scope (ZATCA, Peppol, EN 16931), and buy vs build.
