Website chatbots: when they bring enquiries and when they just annoy
- A bot earns its place where the same questions repeat: order status, stock, opening hours.
- For high-value services with a long cycle it usually gets in the way — people want a person.
- A popup two seconds after arrival lowers conversion, it does not raise it.
- More often it is better to build a proper FAQ and a fast form than a bot.
“Add a bot for us” is a regular request. In roughly half of cases we say no and explain why.
When a bot really works
A store with real volume. “Where is my order”, “do you have size 42”, “how late are you open” — hundreds of identical questions a week. A bot takes the routine off your manager and answers at three in the morning.
Booking with simple logic. A salon, a service centre, a clinic: choosing a specialist, a date and a time works fine without a human.
First-pass qualification. Two or three questions that filter out clearly irrelevant enquiries, so your manager receives a warm contact.
When a bot hurts
A complex service with a long cycle. Someone choosing a contractor for a project worth several thousand dollars does not want to talk to a script. They want to know whether there is a competent person on the other side. The bot is a barrier.
When the answers are not really there. A bot that says “please hold, an operator will join” to four questions out of five is worse than a plain “message us on Telegram” button.
When it appears immediately. A window covering the content two seconds in is the fastest way to get the tab closed.
What it costs
| Option | Timeline | When it fits |
|---|---|---|
| Off-the-shelf service, basic setup | 2–3 days | Testing the idea |
| Scripted bot with CRM integration | 1–2 weeks | Repeating questions, real volume |
| Language-model bot on your knowledge base | 3–5 weeks | Large catalogue, lots of documentation |
The last option sounds appealing but assumes you already have an organised knowledge base. If you do not, that gets built first — and it is a separate project.
Where to start
Open your team’s messages from the last month and write out the twenty most common questions. If fifteen of them are the same, a bot will pay for itself. If they are all different, what you need is a clearer website.
We will read your enquiries and tell you straight whether a bot is needed. If a FAQ and a fast form are enough, we will say so.