Most chatbots are annoying. Here’s what I’m doing differently.

I've spent the last few months building AI chatbot demos for tradies, salons, retail stores and local service businesses. And pretty quickly, one thing became obvious: most chatbot experiences are genuinely bad.

Not because AI is bad. Because most bots are designed around what the business wants — not what the customer needs.

You've probably seen it yourself. You land on a website and within two seconds a chatbot jumps out asking:

• What's your name?
• Your email?
• Phone number?
• How can we help?
• Want to book a call?

Before you've even figured out what the business does. As a customer it's exhausting. And honestly — a bit desperate.

The mistake I keep seeing

A lot of chatbot setups try to force lead capture too early. But real conversations don't work that way. Before someone hands over their contact details, they usually want to:

• Understand what you actually do
• See if you can help with their specific situation
• Get a rough sense of price or process
• Decide whether they trust you

Only then are they comfortable giving details. So I started building flows that reflect that. Instead of opening with “give me your information,” the bot guides the conversation naturally — answers common questions first, reduces friction, and only asks for contact details when it actually makes sense to.

I'm not trying to replace anyone

This is another thing that bothers me about a lot of AI marketing right now — the “replace your staff with AI” angle. That's not what I'm building toward.

Most small businesses don't need a robot pretending to be human. They need:

• Fewer missed enquiries
• Faster initial replies
• Basic qualification done before the follow-up call
• A smoother experience when they're flat-out or after hours

A plumber on site can't answer every call. A salon owner doing treatments can't constantly reply to messages. A retail owner can't sit at the computer answering stock questions all day. That's the gap a chatbot can genuinely fill — not by pretending to be magic, just by being useful.

Staying within boundaries

One thing I've become deliberate about is making sure bots don't overreach. I don't want them giving legal advice, pretending to be qualified engineers, making compliance decisions, or confidently making things up.

The better approach is structured flows, approved information, clear handoff points, and collecting the right details so a real person can follow up properly. That's safer for the business and — genuinely — creates a better experience for the customer too.

What's actually working

The demos that get the best reactions aren't the “smartest” ones. They're the ones that feel fast, simple, helpful, and not pushy. That surprised me a bit at first.

I think small businesses are getting tired of overhyped AI promises. Most just want practical systems that save time, stop missed leads, and make the business look more responsive. That's the direction I'm building toward.

Still figuring it out

I'm still experimenting — different flows, different industries, different integrations, watching how customers actually respond. Some things work better than expected. Some ideas completely flop.

That's part of building something real instead of pretending everything is already solved. And I think more businesses appreciate that kind of honesty anyway.

If you're curious about how a chatbot might actually work for your business, I'm happy to show you a demo or just talk through ideas. No hard sell — just a practical conversation.