How much does a custom AI chatbot cost?

As example ranges, roughly ₹5–12 lakh for a production RAG support bot with integrations and evaluation; ₹1.5–4 lakh for simpler FAQ or lead bots. Messaging channels and enterprise security requirements increase cost.

Is there a monthly fee after the chatbot is built?

Yes -hosting, LLM API usage, messaging BSP fees if applicable, and optional maintenance retainers. Budget ₹3k–75k+ monthly for typical SMB volumes, higher at scale.

AI Chatbot Cost (2026)

AI chatbot pricing for FAQ bots, RAG support assistants, and multi-channel deployments -build costs, monthly run costs, and what proposals should include.

Direct answer

As example ranges, AI chatbot build costs typically fall into: ₹1.5–4 lakh for a bounded FAQ or lead-capture widget with limited integrations; ₹5–12 lakh for a RAG-based support assistant with citations, CRM or helpdesk integration, and evaluation; ₹8–18 lakh for multi-channel (web + messaging) with access control and analytics. Monthly run costs: ₹3k–75k for hosting and APIs at SMB volumes, scaling with traffic and model tier. Sub-₹50k quotes usually omit retrieval, logging, and maintenance.

Price drivers

Channels (web only vs WhatsApp Business API), languages, number of intents, corpus size and refresh, integrations (Freshdesk, Zoho, Shopify), human handoff UX, admin for content updates, and SLA expectations. Customer-facing brand tone and compliance review add design and legal cycles. Each integration is a mini-project -price it explicitly.

Build vs SaaS chatbot platforms

SaaS per-seat or per-conversation pricing wins at small scale with generic FAQs. Custom build wins when answers require private retrieval, complex permissions, deep CRM actions, or the bot lives inside your product. Compare three-year TCO, not month-one subscription. Platform lock-in without export hurts when you outgrow templates.

Messaging channels and localisation

WhatsApp Business API has Meta conversation charges plus BSP fees -model separately from dev cost. Session windows and template messages affect UX design. Mixed-language queries need testing on real user phrasing, not textbook prompts. Teams that expect messaging-first engagement should budget channel work, not only a web widget.

Ongoing costs to budget

Model inference, embedding refresh when docs change, monitoring, prompt updates after product changes, and support retainers (often ₹25k–1.5L/month depending on SLA). Plan regression runs after corpus updates -silent retrieval drift is common. API price changes from OpenAI or Anthropic flow through; contracts should allow model router swaps.

How to compare vendor quotes

Ask for: golden-set accuracy on your data, sample conversation logs with citations, handoff flow demo, data residency statement, and line-item maintenance. Reject quotes that only show a marketing-site bubble with three canned questions. Align with how to build an AI chatbot checklist before signing.

Next step with ZiyadX

Send twenty real support transcripts or lead questions and your channel list. Review /services/ai-solutions and Antomind for a governed assistant reference. Cross-read AI development cost for broader AI budgets. Contact ZiyadX for a fixed-scope vertical slice quote before committing to multi-channel rollout.

Related paths

Frequently asked questions

How much does a custom AI chatbot cost?: As example ranges, roughly ₹5–12 lakh for a production RAG support bot with integrations and evaluation; ₹1.5–4 lakh for simpler FAQ or lead bots. Messaging channels and enterprise security requirements increase cost.
Is there a monthly fee after the chatbot is built?: Yes -hosting, LLM API usage, messaging BSP fees if applicable, and optional maintenance retainers. Budget ₹3k–75k+ monthly for typical SMB volumes, higher at scale.