AI Chatbot for Your Website and WhatsApp: What It Does and How Much It Costs

Neogenio AI

An AI chatbot for your website and WhatsApp automatically answers customer questions in natural language using your company’s information: prices, hours, delivery, services, order status. It can collect quote requests and schedule appointments around the clock, and when it doesn’t know an answer or the customer asks for a person, it hands off the conversation to someone on your team. The cost depends on how much it does: an assistant for frequently asked questions is a small project, while one connected to your store, CRM, and calendar is more complex. Below you’ll find what it can and can’t do, what you need to know about GDPR, and what affects the price.

What is an AI chatbot and how it differs from a traditional one

Traditional chatbots work with buttons and fixed scripts: a customer clicks “Hours” and gets a preset text. If they type something else, the bot doesn’t understand. An AI chatbot uses a language model, so it understands freely worded questions, spelling mistakes, or questions in other languages, and responds naturally.

The important difference is that a good AI chatbot doesn’t make things up. It’s configured to search for answers in the information you give it: your website pages, price lists, delivery terms, internal documents, product database. This technique is usually called knowledge base search or RAG, and it greatly reduces the risk of made-up answers.

The same “brain” can be deployed across multiple channels: a chat widget on your website, WhatsApp, Facebook Messenger, or a contact form. Your customer chooses where to write, and the answers stay consistent.

What an AI chatbot can do on your website and WhatsApp

A well-configured chatbot takes over repetitive work from customer conversations. The most useful features are:

• Answers to frequently asked questions — hours, address, starting prices, delivery zones, returns, warranty.
• Product or service recommendations — asks a few questions and suggests the right option from your offerings.
• Collecting quote requests — gathers the necessary information (what the customer wants, quantity, timeline, contact) and sends it to your CRM or email.
• Scheduling appointments — checks available time slots in your calendar, confirms, and sends reminders.
• Order status — if connected to your online store, tells the customer where their order is.
• Responses in multiple languages — useful if you have customers from abroad.

On WhatsApp, the advantage is that many customers already prefer this channel. Through WhatsApp Business Platform, the chatbot can respond automatically, send confirmations and reminders, and move the conversation to an operator when needed. A chatbot is, in fact, a form of automation; for ideas about other processes you can automate, check out the article on AI automation for online startups.

What it can’t do and where it might fail

It’s important to have realistic expectations. An AI chatbot is an assistant, not an experienced employee.

• It doesn’t know what you haven’t told it. If prices or hours aren’t updated in its knowledge base, it will give outdated information.
• It can make mistakes or “guess.” Even with good configuration, AI models can sometimes give a confident but wrong answer. That’s why you limit its topics and tell it to admit when it doesn’t know.
• It shouldn’t make decisions involving money or legal implications. Special discounts, refunds, complaints, medical or legal advice stay with people.
• It doesn’t handle strong emotions well. An upset customer needs a person, not polite auto-generated text.

The solution isn’t to skip the chatbot, but to set clear boundaries and test real conversations before launch. A set of 30-50 typical questions, checked one by one, quickly shows you where corrections are needed.

Use cases by business type

The table below shows examples for several types of businesses. These are general scenarios that you adapt to your processes.

Notice that each case has a column “When a Person Takes Over.” That’s not a detail, it’s half the project.

Handing off to a person: how handover works

Handover is when the conversation moves from chatbot to a team member. If done poorly, the customer repeats everything from scratch and leaves unhappy. If done well, your colleague sees the full history and continues naturally.

When handover happens

• the customer explicitly asks to speak with a person;
• the chatbot can’t find the answer in its knowledge base;
• the topic is on the sensitive list: complaints, payments, refunds, personal data;
• the conversation exceeds a certain number of messages without resolution.

How it works in practice

The chatbot tells the customer it’s connecting them with a colleague and, outside business hours, clearly states when they’ll get a response. Your team gets a notification via email, Slack, or a shared message inbox, with a summary of the conversation and already-collected data. On WhatsApp, the operator can take over the same conversation, so the customer doesn’t need to switch channels.

Decide from the start who responds, how quickly, and what happens on weekends. A chatbot that promises “a colleague will contact you right away” and then nobody responds does more harm than no chatbot at all.

GDPR and customer data

A chatbot almost always processes personal data: name, phone, email, sometimes address or order details. GDPR applies the same way as for website forms, and a few rules are essential.

• Inform users. Make it clear they’re talking to an AI assistant and where to find your privacy policy. The European AI Regulation (AI Act) also requires people to be informed when they interact with an AI system.
• Collect only what’s necessary. For an appointment you need name and phone, not an ID number.
• Choose vendors with data processing agreements. Check where data is stored, whether it leaves the EU and under what conditions, and whether the AI model vendor uses conversations for training.
• Set a retention period. Conversations shouldn’t be kept forever; define a timeframe and delete automatically.
• Respect people’s rights. You must be able to provide a copy of their data or delete it on request.

On WhatsApp, the customer usually starts the conversation, but for messages you send (reminders, offers) you need their consent, per WhatsApp rules and commercial communication laws. If you’re unsure, discuss the setup with your data protection officer or a legal consultant.

How much an AI chatbot costs

At Neogenio, AI chatbots for websites and WhatsApp are priced per project, based on your needs. We don’t publish a fixed price because an assistant answering 20 frequently asked questions and one connected to your store, CRM, calendar, and WhatsApp are very different projects.

As a general reference, here’s what the market looks like. The values below are an approximate market average, with wide ranges, and vary significantly from vendor to vendor.

What affects the price: number of channels (website, WhatsApp, Messenger), integrations with other apps, conversation volume, number of languages, handover complexity, and data protection requirements. Beyond implementation, there are monthly operating costs: AI model usage, hosting, and on WhatsApp, fees charged by Meta for certain message types. We estimate these together before starting. Details are on our AI solutions and automation pages.

How to launch a chatbot in small steps

1. Gather real questions from emails, phone calls, and WhatsApp from recent months.
2. Write official answers and put them on your website in an FAQ page. They become the chatbot’s knowledge base too.
3. Set boundaries and handover: what topics the bot avoids and who takes over conversations.
4. Launch on one channel, usually your website or WhatsApp, and read conversations in the first few weeks.
5. Correct and expand: add answers, integrations, and other channels based on what customers ask.

If you need something beyond a chatbot, like a customer portal or an internal assistant for your team, read about AI application development or AI automation offered by Neogenio in Bucharest.

Want an AI chatbot on your website or WhatsApp that answers from your company’s information and knows when to call a person? Tell us what questions you get most often, and the Neogenio team will send you a project-based quote.