AI Chatbot for Your Website, Grounded in Your Own Content
An AI chatbot for a website answers visitor questions from content you approve, such as your pages, PDFs and policies, and passes the conversation to a person when it cannot help. I build these with large language model APIs, retrieval (RAG), Python and n8n workflows that send enquiries to your CRM, email or ticketing tool.
By Anurag Srivastav · Delhi NCR, India · Open to remote projects ·
Discuss your project
Custom website chatbot: instant answers grounded in your business content, lead capture and CRM integration.
Who this is for
Teams whose visitors keep asking the same questions, and who have written answers (site pages, brochures, policies) the chatbot can rely on. It works best with a defined set of topics, not an open-ended assistant.
Scope, timeline and pricing
A basic chatbot starts from ₹5,000 INR service fee, confirmed after scoping around your content, the questions visitors ask, where enquiries should go, and who takes over a conversation. Delivery date and final price are confirmed after those are known.
Which kind of chatbot you need
| Type | What it does | Fits when |
|---|---|---|
| Rule-based | Fixed menus and answers | A few predictable questions |
| LLM chatbot | Understands free text, answers from the model's general knowledge | Rarely right for business facts |
| RAG assistant | Retrieves your approved content, then answers from it, with links | Policies, services, FAQs, documentation |
| Agent | Chooses between allowed tools, such as booking or ticket creation | Several dependent actions with approval |
Start with the smallest one that meets the task. Most website chatbots are RAG assistants with a lead-capture step. For multi-step agents, see LLM and AI agent development.
What gets built
- A chat widget for your site
- A knowledge base built from the content you approve
- Answers that link back to the source page
- A fallback message when the content does not support an answer
- A hand-off route to a person
- Enquiry capture into your tools through n8n workflows
- Conversation logs for review
How the work runs
- Agree topics, sources and hand-off rules.
- Build the knowledge base and guardrails.
- Test with real visitor questions and difficult cases before launch.
- Launch with the AI disclosure in place.
- Review logs and correct gaps.
Pricing
| Part | What it covers |
|---|---|
| Service fee | Basic chatbot starts from ₹5,000 INR, confirmed after scoping |
| Third-party costs | Language model API usage, hosting and domain, paid to those providers. They follow each provider's current pricing page. |
| Optional | Extra integrations, additional languages, change requests |
Data, disclosure and the law
Tell visitors they are talking to an AI, at the start of the chat. If the chatbot collects names, emails or phone numbers, India's Digital Personal Data Protection Act and Rules apply to the business running the site. The Rules were notified in November 2025 and the main duties on notice, consent and security start from May 2027 (as of October 2026; the timeline can change). If you serve visitors in the EU, the EU AI Act's chatbot disclosure duty has applied since 2 August 2026.
Limits
A chatbot can still give a wrong answer, even with retrieval. It needs a person to take over, and its content needs updating when your business changes. It depends on third-party APIs that change. No result, such as leads or rankings, can be promised.
Related work from my portfolio
BMRC Hospital AI Feedback Agent
Enterprise AI workflow that processes multilingual patient feedback, runs sentiment analysis, routes tickets automatically and sends Gmail notifications.
Tools: n8n, OpenAI, Supabase, Gmail API
Read the project case studyAI Voice Agent
Real-time conversational voice AI with streaming audio, OpenAI integration and WebSocket communication. The portfolio reports sub-200ms latency, with benchmark methodology not yet published.
Tools: OpenAI, WebSockets, Streaming Audio
Read the project case studyBoth are related work, not website chatbots.
Read my approach to production RAG systems →Questions about website chatbots
Do I need a chatbot or a contact form?
If most questions repeat and the answers already exist in writing, a chatbot helps. If every enquiry is unique, a good form is simpler.
Will the chatbot make things up?
It can. Answering from your approved content, showing sources and saying it does not know reduces the risk but does not remove it.
Can it send enquiries to my CRM or email?
Yes, through n8n workflows or API integrations, after confirming which tool you use.
Is visitor data safe?
That depends on how it is configured and hosted. We agree what is collected, why, and how long it is kept before launch.
What does it cost?
A basic chatbot starts from ₹5,000 INR service fee, confirmed after scoping. Language model usage is billed separately by the provider.
Can it run on WhatsApp?
WhatsApp has its own platform rules and costs. I would check Meta's current documentation before scoping it.
Tell me what you want to build
Share the questions your visitors ask, the content you have, and where enquiries should go. We can discuss scope, constraints and a practical starting point.
Book a project discussionContact Anurag