73% of your website visitors leave without doing anything. A chatbot won't fix all of that — but it can capture leads at 2 AM, answer the same 15 questions your team answers daily, and route serious enquiries to your WhatsApp. Here's how to set one up without writing a line of code.
This guide is written for business owners and founders — not developers. We'll cover what's realistic, what it costs, and how to avoid the common mistakes that make chatbots annoying instead of useful.
What AI Chatbots Actually Are in 2025–2026
Let's clear up the confusion first. There are two fundamentally different types of chatbots:
Rule-Based Chatbots (The Old Way)
These follow scripted decision trees. User clicks "Pricing" → bot shows pricing info. User clicks "Support" → bot asks for order number. It's basically a fancy FAQ page that looks like a chat window.
Pros: Predictable, cheap, easy to set up. Cons: Can't handle anything outside the script. Users get frustrated quickly.
LLM-Powered AI Chatbots (The New Way)
These use large language models (like GPT-4, Claude, or Gemini) to understand natural language and generate responses. You feed them your business data — website content, product info, pricing, policies — and they answer questions conversationally.
Pros: Handle unexpected questions, feel natural, improve over time. Cons: Can hallucinate (make up answers), cost more to run, need guardrails.
The Practical Middle Ground
Most businesses in 2025–2026 use a hybrid approach: an LLM-powered bot for general conversation, with rule-based flows for critical actions (booking, pricing requests, support tickets). The AI handles the conversation; the rules handle the business logic.
Realistic Capabilities for Small Business
Before you set up a chatbot, understand what it can and can't do:
What a Chatbot Can Do Well
- Answer frequently asked questions (hours, pricing, services, location)
- Capture lead information (name, email, phone, what they need)
- Qualify leads by asking the right questions before routing to sales
- Book appointments (via calendar integrations)
- Provide instant responses outside business hours
- Handle multiple conversations simultaneously
- Reduce repetitive queries for your support team by 40–70%
What a Chatbot Can't Do (Yet)
- Replace your sales team for complex, high-value deals
- Handle emotional or sensitive customer complaints well
- Make business decisions (refunds, custom pricing, exceptions)
- Work perfectly without initial setup and ongoing tuning
- Understand context from previous interactions (unless specifically configured)
The Honest ROI
A well-configured chatbot for a small business typically:
- Captures 5–15 additional leads per month that would have otherwise bounced
- Reduces support email volume by 30–50%
- Pays for itself within 2–3 months if you're converting even a fraction of those leads
Platform Comparison
Here's an honest comparison of the platforms we've tested and deployed for clients:
| Platform | Best For | AI Capability | Pricing (Starting) | No-Code Setup | |
|---|---|---|---|---|---|
| Tidio | Small businesses, e-commerce | GPT-powered (Lyro AI) | Free tier; ₹1,500/mo for AI | Yes (paid) | Yes |
| Intercom | SaaS, mid-size companies | Fin AI (GPT-4 based) | ~$39/mo; AI at $0.99/resolution | Yes | Yes |
| ChatBot.com | Businesses wanting visual builder | AI Assist (GPT-based) | ~$52/mo | Via integration | Yes |
| Crisp | Startups, budget-conscious | MagicReply AI | Free tier; ₹1,800/mo pro | Yes | Yes |
| Botpress | Technical teams, custom flows | Built-in LLM | Free tier; pay per usage | Yes | Semi (visual + code) |
| Custom GPT Bot | Unique requirements | Full GPT-4/Claude | API costs (~$5–50/mo) | Via API | No (needs developer) |
| WhatsApp Business API | India, mobile-first audiences | Via BSP providers | ₹5,000–15,000/mo | Native | Partial |
Our Top Picks by Scenario
Just getting started, budget under ₹5,000/month: Tidio or Crisp. Both have free tiers that let you test before committing. Tidio's Lyro AI is genuinely good for small business use cases.
E-commerce store (Shopify/WooCommerce): Tidio integrates natively with both. It can answer product questions, track orders, and capture abandoned cart leads.
SaaS or tech company: Intercom's Fin AI is the gold standard for product support. It's expensive, but if you're handling 100+ support conversations monthly, it pays for itself.
India-focused, mobile-first audience: WhatsApp Business API with an AI layer. Your customers are already on WhatsApp — meet them there.
Custom requirements, developer available: Botpress or a custom GPT bot. Full control, but requires technical maintenance.
Cost Breakdown: What You'll Actually Pay
| Cost Component | Budget Option | Mid-Range | Premium |
|---|---|---|---|
| Platform subscription | ₹0–₹2,000/mo | ₹3,000–₹8,000/mo | ₹15,000–₹50,000/mo |
| AI/LLM usage | Included in platform | ₹500–₹3,000/mo | ₹5,000–₹20,000/mo |
| WhatsApp API | ₹0 (no WhatsApp) | ₹5,000–₹10,000/mo | ₹10,000–₹30,000/mo |
| Setup (one-time) | DIY (₹0) | ₹10,000–₹25,000 | ₹25,000–₹1,00,000 |
| Monthly maintenance | DIY (₹0) | ₹2,000–₹5,000/mo | ₹5,000–₹15,000/mo |
| Total Year 1 | ₹0–₹24,000 | ₹1,00,000–₹3,00,000 | ₹3,00,000–₹10,00,000+ |
Reality check: Most small businesses should start with the budget or mid-range option. You can always upgrade later. Don't spend ₹1 lakh on a chatbot before you've validated that your website gets enough traffic to justify it.
WhatsApp Business API Chatbots
If your business serves customers in India, Southeast Asia, or the Middle East, WhatsApp is not optional — it's where your customers are.
WhatsApp Business App vs WhatsApp Business API
WhatsApp Business App (free): Manual replies, basic auto-replies, catalog. Good for solopreneurs handling under 50 conversations per day. No chatbot capability.
WhatsApp Business API (paid): Programmable, supports chatbots, handles unlimited conversations, integrates with CRMs. Required for any automated chatbot functionality.
How to Set Up a WhatsApp AI Chatbot
- Get API access through a Business Solution Provider (BSP): Gupshup, Wati, Interakt, AiSensy, or Twilio. Indian BSPs like Gupshup and Wati offer competitive pricing.
- Verify your business with Meta (Facebook). You'll need a Facebook Business Manager account and business verification (takes 2–7 days).
- Choose a chatbot builder that integrates with your BSP. Wati has a built-in flow builder. Gupshup works with Botpress and custom solutions.
- Design conversation flows for your top 10 customer queries.
- Train the AI on your business FAQs, product/service info, and pricing.
- Set human handoff rules — when should the bot transfer to a real person?
WhatsApp Chatbot Costs in India
| Provider | Setup | Monthly | Per Conversation |
|---|---|---|---|
| Wati | ₹0 | ₹4,000–₹10,000 | Included in plan |
| Interakt | ₹0 | ₹2,500–₹6,000 | ₹0.50–₹1.00 per conversation |
| AiSensy | ₹0 | ₹2,000–₹5,000 | Per conversation pricing |
| Gupshup | Custom | Custom | ~₹0.50–₹1.50 per message |
Meta charges per-conversation fees on top of BSP costs. Marketing conversations cost more than utility/service conversations. Budget ₹5,000–15,000/month total for a small business handling 500–2,000 conversations.
Training Your Chatbot with Business Data
This is where most chatbots fail or succeed. The AI is only as good as the data you feed it.
What to Feed Your Chatbot
- Your website content — services pages, about page, pricing page, FAQ page. Most platforms can crawl your website automatically.
- Product/service documentation — detailed descriptions, specifications, pricing tiers, terms.
- FAQ list — compile the 20–30 questions your team answers most frequently. Write clear, concise answers for each.
- Pricing information — be specific. "Starting at ₹25,000" is better than "contact us for pricing" (though you can direct complex pricing queries to your sales team).
- Business policies — refund policy, delivery timelines, working hours, service areas.
- Sample conversations — if you have chat logs or email exchanges with customers, these are gold for training.
What NOT to Feed Your Chatbot
- Internal documents with sensitive business information
- Competitor information or comparisons (the bot might repeat them verbatim)
- Outdated pricing or discontinued products
- Anything you wouldn't want a customer to see
Setting Guardrails
Every AI chatbot needs boundaries:
- Define what the bot should NOT answer — legal questions, medical advice, specific competitor comparisons.
- Set a confidence threshold — if the bot isn't sure about an answer, it should say "Let me connect you with our team" instead of guessing.
- Create escalation triggers — keywords like "refund," "complaint," "speak to manager" should immediately route to a human.
- Review conversations weekly for the first month. Look for wrong answers, missed opportunities, and common questions the bot can't handle.
"We set up a chatbot for a D2C skincare brand. The first week, the bot told a customer that their product 'cures acne.' It doesn't — it helps manage acne. That one wrong word could have been a regulatory issue. We added guardrails around health claims immediately. Always review your bot's first 100 conversations carefully."
Common Mistakes (And How to Avoid Them)
Mistake 1: Making the Bot Pretend to Be Human
Don't name your bot "Sarah" and hide the fact that it's AI. Users feel deceived when they realize. Instead, be upfront: "Hi! I'm an AI assistant for [Your Business]. I can help with pricing, services, and booking. For complex questions, I'll connect you with our team."
Mistake 2: No Human Handoff
A chatbot without human escalation is a dead end. When the bot can't help, the user should be able to reach a human within 2 clicks — via WhatsApp, phone, or live chat transfer.
Mistake 3: Too Many Questions Before Value
Don't ask for name, email, phone, company, budget, and timeline before answering a single question. Answer first, capture details naturally within the conversation.
Mistake 4: Set and Forget
Chatbots need ongoing maintenance. New products, updated pricing, changed policies — if you don't update the bot's knowledge, it gives wrong answers. Schedule a monthly review.
Mistake 5: Deploying on a Low-Traffic Site
If your website gets under 500 visitors per month, a chatbot won't generate meaningful ROI. Fix your traffic problem first (SEO, content, ads), then add the chatbot.
Mistake 6: Ignoring Mobile UX
Over 70% of chat interactions happen on mobile. Test your chatbot on a phone. Is the chat window too big? Does it cover important content? Can users easily type and scroll?
Step-by-Step Setup Guide (Using Tidio as Example)
This guide uses Tidio because it has a generous free tier and works well for small businesses. The principles apply to any platform.
Step 1: Create Your Account (5 minutes)
- Go to tidio.com and sign up with your business email.
- Enter your website URL when prompted.
- Choose your industry and company size.
Step 2: Install on Your Website (10 minutes)
For WordPress:
- Install the Tidio plugin from the WordPress plugin directory.
- Activate it and connect your Tidio account.
For Shopify:
- Install Tidio from the Shopify App Store.
- It auto-configures.
For any other website (including Next.js, custom HTML):
- Copy the JavaScript snippet from Tidio's dashboard.
- Paste it before the closing
</body>tag on your website. - For Next.js, add it to your
layout.tsxusing next/script.
Step 3: Configure the Chat Widget (15 minutes)
- Set your brand colours to match your website.
- Write a welcome message. Not "Hi, how can I help you?" Instead: "Hi! Looking for [your main service]? I can help with pricing, timelines, and booking. Ask me anything."
- Set online/offline hours. During business hours: live chat + AI. After hours: AI only with lead capture.
- Add your team photo or logo to the chat widget for trust.
Step 4: Train the AI (30–60 minutes)
- Go to the AI section (Lyro AI in Tidio).
- Add your website URL — it will crawl and index your content.
- Upload additional FAQ pairs — add 15–20 question-and-answer pairs manually for topics your website doesn't cover well.
- Set the bot's personality — professional, friendly, concise. Tell it your business name and what you do.
- Define off-limits topics — what the bot should refuse to answer.
Step 5: Set Up Lead Capture (15 minutes)
- Create a flow: when a user asks about pricing or services, the bot should ask for their name and email before providing detailed info.
- Connect to your email or CRM (Tidio integrates with HubSpot, Mailchimp, Google Sheets, and more).
- Set up email notifications so you're alerted when a new lead comes in.
Step 6: Test Thoroughly (30 minutes)
- Open your website in an incognito window.
- Ask the chatbot 20 different questions — including edge cases.
- Try to break it. Ask irrelevant questions. Ask in Hindi. Ask about competitors.
- Check that lead capture works (test submission → check email/CRM).
- Test on mobile.
Step 7: Launch and Monitor (Ongoing)
- Go live.
- Review all conversations daily for the first week.
- Weekly for the first month.
- Monthly after that.
- Update the bot's knowledge base whenever your business changes.
Chatbot Setup Checklist
- Defined top 20 customer FAQs with clear answers
- Chosen a platform that fits budget and requirements
- Installed chat widget on website
- Customized widget colours and branding
- Written a specific welcome message (not generic)
- Trained AI on website content and FAQs
- Set up lead capture flow with email notifications
- Configured human handoff for complex queries
- Defined off-limits topics and guardrails
- Set online/offline hours with appropriate responses
- Tested on desktop and mobile
- Tested edge cases and irrelevant questions
- Connected to CRM or email for lead notifications
- WhatsApp integration configured (if applicable)
- Scheduled monthly review for bot performance
FAQ
How long does it take to set up a chatbot?
A basic chatbot with AI capabilities takes 2–4 hours to set up using a no-code platform. Training and fine-tuning take another 2–4 hours. Budget a full day for a solid initial setup. Ongoing optimization is about 2–4 hours per month.
Will a chatbot annoy my website visitors?
Only if you implement it badly. Don't auto-open the chat on every page. Don't pop up within the first 3 seconds. Do make it available but not intrusive. A small icon in the corner that visitors can click when they want help is the right approach. Let the user initiate.
Should I use a chatbot or live chat?
Both. Use AI for first response and common questions. Route to live chat (human) for complex queries, complaints, and high-value leads. This hybrid approach gives you 24/7 availability without 24/7 staffing costs.
How do I measure chatbot ROI?
Track these metrics monthly: (1) Number of conversations initiated, (2) Lead capture rate (conversations that capture contact info), (3) Human handoff rate (should decrease over time), (4) Customer satisfaction rating (most platforms include this), (5) Support ticket reduction (compare before/after). If your chatbot captures 10 leads per month and you convert 20% at an average deal value of ₹50,000, that's ₹1,00,000 in revenue from a ₹5,000/month tool.
Can a chatbot work in Hindi or other Indian languages?
Yes. Most LLM-powered chatbots (GPT-4, Claude) handle Hindi, Hinglish, and other major Indian languages well. Tidio's Lyro and Intercom's Fin support multilingual conversations. WhatsApp chatbots via Gupshup and Wati also support multiple languages. Test thoroughly — AI quality varies by language.
Do I need a developer to set up a chatbot?
For basic setups on WordPress, Shopify, or Wix — no. The platforms provide plugins or embed codes that anyone can install. For custom websites, Next.js apps, or complex integrations (CRM, WhatsApp API, custom flows) — you'll likely need a developer for the initial setup, but not for ongoing management.
What happens when the chatbot can't answer a question?
A well-configured bot should: (1) Acknowledge it doesn't know the answer, (2) Offer to connect the user with a human, (3) Capture the user's contact info so your team can follow up, (4) Log the unanswered question so you can add it to the knowledge base. Never let the bot guess or make up an answer.
What to Do Next
Start small. Pick one platform from the comparison above, set up a basic chatbot with your top 15–20 FAQs, and run it for 30 days. Measure the results before investing more.
If you want it done right from day one, we set up and train AI chatbots for businesses — starting at ₹25,000. That includes platform selection, AI training on your business data, lead capture flows, and WhatsApp integration.