A chatbot answers. A receptionist books.




They were built for different businesses
Chatbots were designed by software companies with a support problem: thousands of tickets a week, most of them answerable from documentation. Success is a question answered without a human. Deflection is the metric, and for that business it is the right one.
A local business has the opposite problem. You do not have too many enquiries. You have too few, and the ones you get are precious. Answering a question and ending the conversation is not success — it is a missed booking with extra steps.
Three differences that decide whether it pays for itself
Where it lives. A chatbot is on your website. Your enquiries are not — they are on the phone, in texts, in social messages, in email. A tool that only covers the website covers a minority of your enquiries while feeling like it covers them all.
What it does at the end. A chatbot answers and stops. At best it links to a booking page, which is a second decision and a second chance to leave. The useful version offers a time and confirms it inside the same conversation.
What it does when it does not know. This is the one that matters most. Many chatbots will improvise, confidently, and a business has then been committed to something in writing by software. The correct behaviour is to say so, hand over, and record the gap so somebody can write the answer once.
What to ask before buying either
- Can it answer the phone, or only the website?
- Does it write into my actual calendar, or link to a booking page?
- What does it do when it does not know the answer — exactly?
- Can I see every conversation it had, including the ones it got wrong?
- When a person takes over, do they see the whole history?
The fourth is the one vendors least like answering and the one that tells you the most. A system you cannot audit is one you cannot trust with customers.
When a chatbot is the right answer
If you have a large documentation set, a support team already in place, and the goal genuinely is to reduce ticket volume, a chatbot is the correct tool and this article does not apply to you. The failure is not chatbots. It is buying a deflection tool to do an acquisition job.