The one question I ask before building anything
When somebody tells me they want their phone answered, the useless question is “what should it say?”
Nobody knows. They haven’t thought about it in those terms, and if you push, you get a generic script back — hours, address, thanks for calling. That script is worth almost nothing, because it only handles the calls that were never in danger.
So I ask something else.
“Tell me about your worst phone call.”
Not the busiest one. The worst one. The one that wasted forty minutes, or blew up into a complaint, or ended with somebody hanging up on you.
People remember these in detail. You get the story straight away, with the actual words in it. And inside that story is everything I need:
- What the customer wanted that the business couldn’t give them
- Where it went sideways — a price, a policy, a promise somebody made
- What the owner wishes had happened instead
That last one is the real prize. Ask “what does a good call sound like” and you get a shrug. Ask “what should have happened on that one” and you get a specific, testable answer.
Bad calls tell you where the edges are
A good call is easy and forgettable. It tells you nothing, because anything would have handled it.
Bad calls are where the rules live. The tenant who thinks a fee applies on the wrong day. The person who wants a service you stopped offering. The one who’s already been quoted something over the phone by somebody who shouldn’t have quoted it. The one who’s furious before they say hello.
Every one of those is a rule that exists in the owner’s head and has never been written down anywhere. That’s exactly the material I need, and asking directly — “what are your rules?” — never surfaces it. Asking about a bad night does.
It also tells you what should never be automated
Half the answers I get are some version of that one should have come straight to me.
Good. That’s a boundary, and it’s more valuable than any script. A phone that knows which calls to hand over is worth more than one that tries to handle everything, because the calls it shouldn’t touch are the ones that cost real money when they go wrong.
Emergencies get a person. Angry gets a person. Anything with a number attached that isn’t already written down gets a person. Those lines get drawn from bad-call stories, not from a features list.
Try it on yourself
If you run something, do this for your own business tonight. Write down the last three calls that ruined an hour of your day. Then write what should have happened on each one.
You’ll have a better specification for your front desk than most businesses ever write down — whether or not anybody ever automates it.
That’s the first conversation I have with anybody. If you want to have it, tell me what’s breaking.