Chatbot quotes are unusual in one respect: the build cost is often the smaller number. A rule-based bot costs a fraction of an LLM bot to build and almost nothing to run. An LLM bot costs more to build and then charges you for every conversation, forever. Choosing on build price alone is how businesses end up with an annual bill larger than the project.
What each type costs to build
| Type | Build cost | Timeline | Handles |
|---|---|---|---|
| Rule-based FAQ bot | ₹35,000 – ₹1.25 lakh | 2 – 3 weeks | 20 – 50 predefined questions |
| Intent-based NLP bot | ₹1.25 lakh – ₹3 lakh | 4 – 6 weeks | 50 – 200 phrasings |
| LLM bot with your knowledge | ₹2.5 lakh – ₹6 lakh | 6 – 10 weeks | Open-ended questions |
The right choice is decided by your actual customer questions, not by which technology sounds most current. If eighty per cent of your enquiries cluster around twenty common questions, a rule-based system deflects most of them efficiently and — importantly — cannot invent an answer. If your customers ask varied, unpredictable questions about your own documents, that is what a retrieval-based LLM approach is genuinely for.
The monthly cost is the real decision
A rule-based bot costs ₹500 to ₹2,000 a month in hosting and nothing per conversation. An LLM bot charges per interaction, and the model tier changes that number by an order of magnitude.
| Model tier | Approx. cost per conversation |
|---|---|
| Small models | ₹80 – ₹200 |
| GPT-4o class | ₹400 – ₹1,200 |
| Claude Sonnet class | ₹500 – ₹1,500 |
Multiplied out, the difference becomes the whole business case:
| Monthly conversations | Monthly API cost |
|---|---|
| 500 | ₹500 – ₹3,500 |
| 5,000 | ₹4,000 – ₹25,000 |
| 25,000 | ₹18,000 – ₹1,00,000 |
At high volume, a frontier model's annual API cost can exceed the entire build investment within twelve months. Calculate this before selecting a tier, not after the first quarterly bill.
Cutting the running cost by 60 to 80 per cent
Three techniques reduce LLM cost substantially without a noticeable drop in quality:
- Cache common responses so identical questions are not paid for twice
- Limit the context sent with each request rather than passing everything available
- Route simple queries to a cheaper model and reserve the expensive one for genuinely hard questions
Together these typically take sixty to eighty per cent out of the bill. They are engineering decisions made during the build, and retrofitting them later is more work than including them from the start — so ask any supplier how they intend to handle each of the three.
WhatsApp changes the arithmetic
WhatsApp integration adds ₹5,000 to ₹20,000 a month, plus Meta's own per-message fees. It also brings the twenty-four hour service window, which limits when you can message a customer who has not recently messaged you. For most Indian businesses WhatsApp is where the customers already are, so this is usually worth paying — but it should be a decision made with the numbers in front of you.
Budget for the second year
Chatbots degrade. Customer questions evolve, products change, and reported accuracy loss sits in the region of ten to fifteen per cent a year without review. A neglected bot performs worse after eighteen months than it did at launch, which is the worst possible outcome — you are still paying for it and it is now damaging the customer experience it was bought to improve.
A monthly review cycle — updating the knowledge base, refining prompts, checking what the bot got wrong — keeps it useful. Any proposal that stops at launch is describing half the project.
If you want help working out which of the three types your enquiry volume actually justifies, that is a conversation we are happy to have before any commitment.
Frequently asked questions
How much does an AI chatbot cost in India? +
A rule-based FAQ bot costs ₹35,000 to ₹1.25 lakh over two to three weeks. An intent-based NLP bot runs ₹1.25 lakh to ₹3 lakh over four to six weeks. An LLM bot trained on your own documents costs ₹2.5 lakh to ₹6 lakh over six to ten weeks.
What does a chatbot cost to run each month? +
A rule-based bot costs ₹500 to ₹2,000 a month in hosting with nothing per conversation. An LLM bot charges per interaction — at 5,000 conversations a month that is roughly ₹4,000 to ₹25,000, and at 25,000 conversations it can reach ₹1 lakh.
Which type of chatbot do I actually need? +
If eighty per cent of your enquiries cluster around twenty common questions, a rule-based bot handles them efficiently and cannot invent an answer. If customers ask varied, unpredictable questions about your own documents, an LLM approach is what that is genuinely for.
How do I reduce chatbot API costs? +
Caching common responses, limiting the context sent with each request, and routing simple queries to a cheaper model typically cut sixty to eighty per cent from the bill without a noticeable drop in quality. These are build-time decisions, so raise them before development starts.
What does WhatsApp integration add? +
₹5,000 to ₹20,000 a month, plus Meta's own per-message fees. It also brings the twenty-four hour service window, which limits when you can message a customer who has not recently messaged you.
Do chatbots need ongoing maintenance? +
Yes. Accuracy degrades in the region of ten to fifteen per cent a year without review, as customer questions and your own products change. A neglected bot performs worse after eighteen months than it did at launch.
