Website Chatbots: Do They Really Work or Just Hype?

Website Chatbots: Do They Really Work or Just Hype?

In 2024 and 2025, the chatbot became the accessory everyone wanted to add to their site. Web agencies selling it as a solution to everything, AI platforms promising infallible virtual assistants, tutorials on how to build your chatbot in ten minutes.

Now that the dust has settled a bit, we can make a more honest assessment: when does a chatbot really make sense, and when is it just noise?

The Difference Between a Useful and Useless Chatbot

A useful chatbot answers real questions your visitors ask frequently, accurately, without wasting time. It reduces the number of emails and calls you get for simple things. It works at 11 PM too.

A useless chatbot says "I didn't understand your question" to half the messages, responds generically to everything, or exists only because "competitors have one too". Worse: it makes your site look unprofessional.

The deciding factor isn't the technology. It's whether you actually have a volume of repetitive and predictable requests worth automating.

When a Chatbot Makes Sense

It makes sense if you regularly get the same questions: hours, prices, how a service works, delivery times, geographic coverage. If you answer the same three questions ten times a week, a well-configured chatbot gets them off your back.

It makes sense if your site gets decent traffic volume. On a site with 200 visits per month, a chatbot is hard to justify. On a site with 2,000 visits per month and even a small percentage of inquiries, it starts to have value.

It makes sense for e-commerce with large catalogs, medical or legal offices with many preliminary questions, services with clear qualification processes (for example: "Do you also work in Milan? Do you install this type of system?").

It doesn't make sense for a craftsman with few clients a month, for professionals where every request is different, or for activities where human contact is part of the service itself.

Real Costs

This is where many ideas become clear.

The simplest chatbots — based on predefined flows, like Tidio or Crisp in the basic version — cost 0-30 euros per month. They work well for simple FAQs, but don't handle complex questions.

Real AI chatbots, based on models like GPT-4, cost more. You can use services like Intercom, Drift, or build something custom with OpenAI APIs. Plans with advanced AI start from 50-100 euros per month and climb quickly with volume.

Then there's the hidden cost: configuring them properly. A working chatbot needs an accurate knowledge base — answers to frequently asked questions written clearly, edge cases handled, cases where it needs to escalate to a human. This requires initial work and periodic maintenance.

A poorly configured chatbot gives wrong answers. And a wrong answer to a potential customer is worse than silence.

Simpler Alternatives for Beginners

If you want to handle out-of-hours requests without a full chatbot, you have lighter options.

Contact form with auto-reply: whoever writes gets an immediate email confirming receipt and indicating response times. Simple, works, breaks nothing.

Well-done FAQ on your page: a curated frequently asked questions section reduces inquiries with no additional cost. Often it's the first thing missing.

Simple live chat widget: tools like Crisp or Tawk.to in the free version show a chat, but if you're not online nobody responds — the visitor leaves a message and you reply when you can. Less powerful than an AI chatbot, but less risky.

These solutions aren't inferior. They're adequate to the size of the problem.

How to Evaluate If It Makes Sense for You

Ask yourself these questions.

How many requests do you get per week that are basically the same? If the answer is "few", the chatbot won't change your life.

Have you already written the answers to those questions somewhere? A chatbot without good content to draw from doesn't work.

Are you willing to configure it and keep it updated? If you install it and forget about it, better not to put it in.

If the answers point you toward "yes", it's worth evaluating. If not, invest that budget in something that brings more traffic to your site — it probably pays off more.

If you want to understand if a chatbot makes sense for your specific site, get in touch: I'll tell you what I recommend based on your business type and volume.