Chatbot development cost in India: the two budgets you must plan
Every estimate of chatbot development cost in India has a one-time build cost and an ongoing running cost, and quotes that show only one of them are incomplete. Plan both before comparing vendors.
The build covers design of the conversation, the bot logic, integrations with your CRM, store or database, testing, and launch. With us it starts from ₹40,000. The running cost covers what the bot consumes every month: AI model tokens if it uses AI, Meta’s template message charges if it runs on WhatsApp, hosting, and maintenance once any free period ends.
The two budgets pull against each other. A cheaper build that sends every question to a large AI model can cost more each month than a slightly pricier build that answers common questions with buttons and uses AI only for the rest. A bot designed so customers start the conversation keeps WhatsApp charges low. Good design is how you control chatbot development cost in India over years, not just at launch.
- One-time: design, build, integrations, testing, launch, from ₹40,000.
- Monthly, usage-based: AI tokens, WhatsApp template messages.
- Monthly, fixed-ish: hosting and database.
- Optional: maintenance after 2 free months, from ₹8,000/mo.
How much does each type of chatbot cost to build in India?
On chatbot development cost in India by type: rule-based bots sit at the low end of the AI automation band from ₹40,000, AI bots on company data in the middle to upper range, and bots with staff dashboards or deep integrations from ₹60,000. The type sets the effort far more than the channel.
A menu bot is a decision tree: few states, predictable testing, no AI tuning. A lead bot adds question validation and a write to your CRM. An order or booking bot adds carts, slots, availability checks and confirmations, each with failure cases to handle. An AI bot on your documents adds ingestion, search, answer rules, citations and a test set of real questions. An action-taking agent adds tool design, permissions and far more testing.
Integrations are often the hidden driver. Writing a lead into a Google Sheet is quick; syncing orders with an ERP whose API is poorly documented can take longer than the bot itself. Languages add writing and testing time. A staff inbox where humans take over chats is effectively a small web application.
We quote each of these as a separate line, so you can see exactly what moves your chatbot development cost in India and drop or defer items to fit your budget.
How to estimate AI model cost per conversation
For AI bots, the running side of chatbot development cost in India comes down to tokens. AI model cost per conversation is the number of input tokens times the input rate, plus output tokens times the output rate, summed across every AI reply in the chat. Providers publish rates per million tokens, so the arithmetic is simple once you know your token counts.
Input tokens include much more than the customer’s question: the system instructions, the chat history so far, and any document passages retrieved to answer. For a bot answering from company data, those passages usually dominate. Output tokens are the reply itself, and short replies are both cheaper and better on a phone screen.
A practical estimate: count average AI replies per conversation, estimate tokens per reply from a pilot, multiply by expected monthly conversations, and apply your chosen model’s published rates. Then add a margin for longer chats and retries. We run this calculation with you using rates from the provider’s current pricing page, because rates change and differ widely between models.
- Per reply: (input tokens × input rate + output tokens × output rate) ÷ 1,000,000.
- Per conversation: per-reply cost × AI replies in a typical chat.
- Per month: per-conversation cost × expected conversations, plus a safety margin.
Levers that cut it: answer common questions with buttons (zero tokens), send fewer and shorter passages, trim old chat history, use a smaller model for simple questions, and cache frequent answers. For non-urgent work such as nightly summaries or bulk document indexing, OpenAI’s documentation describes a Batch API at a 50% discount with results within 24 hours.
What WhatsApp message charges add to chatbot running costs
If your bot runs on WhatsApp, Meta charges per delivered template message by category and recipient country, and non-template replies inside an open customer service window are free. That makes the WhatsApp part of chatbot development cost in India depend heavily on who starts the conversation.
Meta’s developer documentation says per-message pricing replaced conversation-based pricing on 1 July 2025. When a customer messages you, a 24-hour customer service window opens in which your bot’s replies are free. Chats that start from a Click-to-WhatsApp ad or a call-to-action button can open a 72-hour free entry point window. Marketing, utility and authentication templates are charged, with volume tiers for utility and authentication.
For Indian businesses there is a billing change to plan for: Meta’s documentation says rupee billing launched on 1 January 2026, and eligible accounts must migrate to INR by 31 December 2026 or face message delivery suspension from 1 January 2027. Check Meta’s pricing page for current rates when budgeting.
A website chat widget has no per-message fee at all, which is one reason some businesses start there. Our WhatsApp Business API cost in India page explains the rate card in detail.
Is a SaaS chatbot cheaper than a custom chatbot?
On chatbot development cost in India over several years, a SaaS chatbot is cheaper to start and a custom build is often cheaper to own after the first year or two, especially once you need integrations the tool lacks or your plan tier climbs with usage. Compare them over 36 months, not one invoice.
SaaS tools charge a recurring fee, usually tiered by contacts, conversations, agents or features. They are a sensible choice for testing whether customers will use a bot, and for simple menus. The costs creep through tier upgrades, add-ons for AI replies or integrations, and sometimes a margin on WhatsApp messages.
A custom bot has a larger upfront cost and no software subscription afterwards. You pay usage and hosting directly and choose whether to take a maintenance plan. It also avoids the cost of leaving: when flows live in a SaaS tool, moving means rebuilding them.
To compare fairly, write both options into the same lines: setup, 36 months of software fees at the tier you will realistically need, usage with any markup, integrations, and the exit cost. The sample quotes and table on this page give you a template. Our custom chatbot vs ChatGPT page covers the related “why not just use ChatGPT?” question.
What increases chatbot development cost in India?
Integrations, channels, languages, document mess and human hand-off features raise chatbot development cost in India more than the choice of AI model. Knowing the drivers lets you trim scope before asking for quotes.
- Each integration: CRM, ERP, store, accounting or booking software, especially with weak APIs.
- Each channel: website, WhatsApp, Instagram, app; each has its own rules and testing.
- Messy knowledge: scanned PDFs, complex tables and conflicting document versions.
- Languages: Hindi, Hinglish or regional flows need writing, approval and testing.
- Transactions: payments, refunds, bookings and partial states multiply test cases.
- Staff tools: inboxes, assignment, reports and admin panels are a web app in themselves.
- Access rules: different answers for different users or dealers.
- Compliance needs: consent capture, audit logs, data retention and deletion workflows.
The model itself is rarely the big number in the build. It is the plumbing around it that takes the time.
How to reduce chatbot cost without making a worse bot
To cut chatbot development cost in India, phase the scope, answer the top questions with buttons, start on one channel and reuse content you already have. These cut both the build and the monthly bill without hurting customers.
Phase one should do one job well, usually lead capture or the ten most common questions. Add ordering, payments or a second channel once the first shows results. Each phase is quoted separately, so you never pay for features nobody uses.
Buttons for predictable tasks cost nothing per chat and are faster for customers than typing. Keep AI for free-text questions that buttons cannot handle. Your existing FAQ page, brochures and staff WhatsApp replies are the raw material; the less new content we need to write with you, the shorter the build.
Finally, choose integrations carefully. A daily export to a Google Sheet may be enough for a year before a live CRM sync is worth paying for. We will suggest cheaper routes in the estimate where they genuinely fit.
Sample quote 1: a rule-based WhatsApp lead bot for a coaching institute
This sample shows how we lay out chatbot development cost in India in a quote; it is illustrative, not a real client. Picture a coaching institute that wants parents to pick a course, batch timing and branch on WhatsApp, with every enquiry landing in a Google Sheet and a counsellor alert.
The build is a button-and-list flow in Hindi and English with a hand-off to a counsellor. It sits at the lower end of the AI automation band from ₹40,000. There are no AI token costs because there is no free-text AI. Meta charges apply only when the institute sends templates, such as a demo-class reminder the next day outside the 24-hour window; replies to parents who message first are free.
- Line 1, build: flows, sheet write, counsellor alert, two languages; from ₹40,000.
- Line 2, Meta setup: business verification, number and display name, done with you on a screen share.
- Line 3, usage: Meta template charges billed by Meta to the institute.
- Line 4, hosting: small serverless setup in the institute’s cloud account.
- Line 5, upkeep: 2 months free, then optional from ₹8,000/mo.
Sample quote 2: an AI website chatbot trained on company data
Again illustrative: a machinery supplier wants its website to answer technical and warranty questions from about two hundred product PDFs, with each answer showing its source and unanswered questions passed to the sales team by email.
This is a retrieval-based AI bot: documents are parsed, chunked and indexed; each question retrieves passages, and a model answers from them with citations. The build sits in the middle to upper part of the band from ₹40,000, depending on how many PDFs are scans and how complex the tables are. Running cost is mostly model tokens per answer, estimated from a pilot using the formula above, plus a small database and server.
- Line 1, build: ingestion, search, answer rules, citations, widget, email hand-off; from ₹40,000.
- Line 2, evaluation: test set of real customer questions, scored before launch.
- Line 3, usage: embedding and answer tokens billed by the model provider to the supplier.
- Line 4, hosting: Postgres with vector search and an app server in the supplier’s cloud.
- Optional line: admin panel to upload documents and review gaps, from ₹60,000.
Sample quote 3: a multi-channel bot with a staff inbox
Illustrative once more: a chain of four diagnostic centres wants one bot on WhatsApp and the website for test prices, home-collection booking and report status, with staff taking over chats in a shared inbox that shows which branch handles what.
This combines a transactional bot (bookings, report lookups from the lab system), an AI layer for free-text questions about preparation and prices, and a staff inbox with assignment and reports. The inbox makes it a custom web app, so the quote starts from ₹60,000. Running costs include Meta template charges for booking confirmations and report-ready alerts sent outside the service window, AI tokens for free-text replies, and hosting.
- Line 1, bot build: booking, report status, AI answers, two channels; within the AI automation band from ₹40,000.
- Line 2, staff inbox: assignment, notes, branch routing, reports; from ₹60,000.
- Line 3, integration: lab system API or database access, scoped after reviewing its documentation.
- Line 4, usage: Meta and model charges billed directly to the centres.
- Line 5, compliance support: consent capture, access controls and audit logs for patient data.
Hidden chatbot costs to ask about before you sign
The hidden parts of chatbot development cost in India are usually usage markups, tier jumps, per-seat fees for your staff, paid integrations, overage charges and the price of leaving. Ask about each in writing.
Useful questions for any vendor: Who pays Meta and the AI provider, and at exactly what rate? Is any margin added? What happens to my monthly fee when contacts or chats double? Are staff logins charged per seat? Which integrations are included and which are extra? Is there a limit on AI replies before overage applies? Can I export my flows, chat history and contacts if I leave, and in what format?
Also ask what is not included: content writing, translation, Meta business verification support, test data preparation and training for your staff. A quote that looks lower can simply leave these to you.
On our side, usage goes to your accounts without margin, there are no per-seat fees in a custom build, and the code is yours. Anything not in the itemised quote is not billed until you approve a written change.
Why do chatbot quotes in India vary so widely?
Quotes for chatbot development cost in India vary widely because vendors price different things: some include integrations, testing and ownership, others only a template bot; some bundle usage with a margin, others pass it through; and team size and overheads differ. The number alone tells you little.
Marketplaces such as Upwork, Fiverr and Freelancer.com list chatbot gigs across a huge range, often for template bots on a vendor’s own platform. Larger agencies price in project managers, designers and account staff. SaaS tools spread the cost into monthly fees. None is wrong; they are different products.
To compare, ask each quote to state: the bot type and flows, every integration, channels, languages, what testing happens, who owns code and accounts, how usage is billed, and what upkeep costs after launch. Our hire an AI developer guide lists questions for vetting the person doing the work.
We are a small freelance team of three, which keeps overheads low and puts you in direct contact with the people building your bot. Our limits are just as clear: we do not staff a 24/7 support desk or run huge programmes needing twenty developers.
GST, payment and invoicing for chatbot projects in India
Chatbot projects in India are paid by UPI or bank transfer against a GST invoice, with the amount and milestones set in the written quote you approve. Nothing is billed before that approval.
Usage bills arrive separately from each provider. Meta bills WhatsApp messages to the payment method on your WhatsApp Business Account; AI providers bill your card or account; your cloud provider bills hosting. Keeping these in your name makes accounting cleaner and means your bot keeps working regardless of any vendor relationship.
International clients are quoted in USD, from US$600 for a bot build, and pay by Wise, bank wire or PayPal. For how GST applies to your own business, please ask your accountant; we do not advise on tax. Terms are set out in your quote and on our terms page.
How long does chatbot development take, and does speed change the cost?
Time and chatbot development cost in India are linked. Most chatbots take 2–4 weeks to build; simple menu bots can be ready sooner, and multi-channel bots with staff inboxes take longer. Squeezing the timeline mainly cuts testing, which usually costs more later than it saves now.
External steps often set the pace. Meta business verification and template approvals, access to your CRM or ERP, and your team’s time to review flows and wording all sit on the critical path. Starting them on day one keeps the build on schedule.
A realistic plan: week one for flows, content and account setup; week two for the build and integrations; week three for testing with your staff on real phones; week four for launch and fixes. AI bots add an evaluation round with real questions. If a deadline such as an admission season or festival sale is fixed, tell us at the start so scope can be phased to meet it.
What does a chatbot cost to maintain each year?
Maintenance is the long tail of chatbot development cost in India. It covers API version updates, flow and content changes, integration fixes when connected software changes, and monitoring. Our builds include 2 months of free maintenance after launch, then optional plans from ₹8,000/mo, or changes quoted one at a time.
Meta and AI providers update their APIs and retire old versions, so bots need occasional technical updates. Your business changes too: prices, offers, branches, timings and policies. AI bots also benefit from a monthly look at unanswered or down-voted questions, which often reveals a missing document or a confusing flow.
Because code and accounts are yours, you can also give maintenance to your own developer. Our maintenance page explains how we handle ongoing work across projects.
Chatbot budget checklist before you ask for quotes
Answer these questions first and every quote for chatbot development cost in India you receive will be easier to compare, and usually lower, because vendors have less to guess.
- The one job the bot must do well in phase one.
- Channels: website, WhatsApp, app, or more than one.
- Rule-based buttons, AI answers, or a mix.
- Languages customers actually use.
- Systems it must read from or write to.
- Expected conversations per month, now and in a year.
- Who takes over when the bot cannot help, and when.
- Documents or FAQs that hold the answers.
- Who will own the WhatsApp, AI and cloud accounts.
- Budget for the build, and a separate monthly ceiling for usage.
Send the list to us on WhatsApp and the estimate will separate build, usage and hosting line by line.
Chatbot development across India
Chatbot development cost in India is the same with us wherever you are, since we work remotely with every client, in English and Hindi, and bill by UPI or bank transfer against a GST invoice.
Enquiries reach us from retailers, clinics, schools, distributors, real estate firms and D2C brands in Mumbai, Delhi, Bengaluru, Kolkata, Jaipur, Surat, Bhopal, Visakhapatnam, Kozhikode and Ranchi. Local languages and industries shape the flows, but the two-budget method applies everywhere.