You finish walking someone through what they need to do. You ask if that makes sense. They say yes.
Three weeks later they haven't done it, and when you ask what happened they're vague. You think back to the call and remember them saying yes. They did say yes. You heard it.
They were never confident. They just weren't going to say so in front of six colleagues with four minutes left in the meeting.
That's the uncertain yes, and it's the quietest way a customer relationship goes wrong. Nothing looks broken. There's no complaint, no escalation, no bad survey. There's a person who agreed to something they didn't understand, and a vendor who wrote it down as understood and moved on.
Why people give it
Two reasons, and both are ordinary.
Nobody wants to look like the one who didn't follow. That gets worse with an audience, and training sessions almost always have an audience.
And a question costs time. If the meeting is nearly over, asking means keeping everyone late so you can admit you're lost. Most people take the yes and plan to figure it out later. Later doesn't come.
You've done this. I've done it inside the last month.
The instrument
The fix isn't asking harder. It's asking something that can't be answered with a yes.
Don't ask a yes or no question. Instead of does that make sense, ask this:
Of the things I just laid out, what do you have concerns about?
That forces a sentence. There's no version of that question you can nod through.
When they say no concerns, close the escape hatch. Most people will still say they're fine. Take it seriously and then make it expensive:
Just to be absolutely clear, when you say you're good, that means you're fully confident in your team's ability to execute on the things we just talked about.
Watch what happens. A real yes stays a yes. An uncertain one usually cracks right here, and what comes out is specific. I don't know how we'd do this on a daily basis. That's the thing you needed.
If they hold, accept it and hand it back. You can't interrogate someone into confidence they don't have, and after two honest passes you have to take the answer:
Alright, I'm going to hold you to that. But if it changes, you've got to let me know.
Say it with a smile. It does two things at once. It puts a small weight on the commitment, and it gives them a door to come back through later without admitting they were wrong on the call.
On confidence scores, and when not to ask
Scoring confidence one to ten is a good instrument used at the wrong moment constantly.
If you've just demonstrated something, the customer hasn't touched it. They watched you do it. Asking how confident they are is asking them to rate a thing they haven't attempted, and the honest answer is that they don't know yet. You'll get a number, and the number will mean nothing.
So delay the ask. Show them, let them have contact with it, then score it on the next conversation when there's something real to rate. After that the score is worth having, because you're watching it move. A seven that becomes a nine is information. A seven collected before they logged in is noise.
And some days you shouldn't ask at all. If the session went sideways for reasons that have nothing to do with them, you already know what the number is, and collecting it teaches the customer that the question is a formality.
Build for rules. Expect exceptions. When you hit one, say you hit one and say why, rather than running the play because it's on the sheet.
Why this matters more than the data
The score was never the point.
The point is that you never want a customer able to ask you why did you let me say yes to that.
Nobody says it in those words. What they do instead is disengage, quietly, from something they never understood, and you find out a quarter later when the usage report is flat. They agreed. You proceeded. Both of you were being polite, and being polite cost them the outcome they paid for.
Asking one uncomfortable question at the end of a call is cheap. Finding out in month four is not.
The uncertain yes is one instance of a wider habit, which is taking what somebody says at face value when the words and the meaning have come apart. That's a listening problem before it's a question problem, and the fix is the same in both directions: run what you hear through a filter rather than recording it accurately.
