How to test whether your AI is losing you enquiries
Most AI enquiry systems are set up once and not tested again. These are the ten messages to send to your own business to find out whether yours is working.
The ten test messages
Send each of these as if you are a new prospect contacting the business for the first time:
1. "Are you available this week?"
2. "How much do you charge for [your main service]?"
3. "Do you cover [an area ten miles outside your standard zone]?"
4. "What happens if I need to cancel a booking?"
5. "I have a specific requirement — [describe an unusual but realistic job type]."
6. "A friend recommended you. What's the next step?"
7. "I left a message yesterday and nobody got back to me."
8. "What's your earliest available slot?"
9. "Do you offer a guarantee?"
10. "I need someone urgently — can you help today?"
How to score each response
A response passes if it answers the question directly, sets a clear next step, and does not make a claim the business cannot honour — such as a specific availability that may not exist.
A response fails if it is a generic holding reply with no specific content, gives an incorrect or outdated answer, states availability without having access to the diary, or has no handoff protocol when the prospect clearly wants to speak to a person.
The three failures that appear most often
One: the AI gives a pricing answer that does not match the current rates. Usually caused by outdated information in the original setup.
Two: the AI confirms availability without diary access. It says "this week" when the next appointment is three weeks out.
Three: the AI has no handoff trigger. The prospect signals clearly that they want a person, and the AI continues trying to resolve the query itself.
Record your scores. A fail on tests 2, 8, or 10 is urgent. The others can be corrected in order of frequency.
EveryCatch configures AI that passes this audit before it handles a single live enquiry. Book a free discovery call to see what a properly configured setup looks like.
