What selling tickets at a beer festival taught me about AI
An experiment at the counter revealed more than expected — about people, indecision and what AI is actually being built for.
AI Loods · Reading time: 4 minutes · Observation & reflection
For the seventh year running, I volunteered at a beer festival, at the entrance where visitors buy their tickets and tokens. After six editions, I knew the rhythm by heart. So this year, to keep things interesting, I ran a small experiment.
I said hello and smiled. Then I waited.
Instead of steering the conversation — making suggestions or filling the silence — I greeted everyone and simply waited for them to tell me what they wanted.
The responses fell into three groups.
Three kinds of visitor
-
Some people knew exactly what they wanted. A ticket plus ten tokens. Keep the change. Done. No hesitation, no doubt. They had thought about it before joining the queue.
-
Then there were people who handed me fifty euros and looked at me expectantly. No request. No indication of what they wanted in return. Just money on the counter and a kind of open waiting — as though the transaction would work itself out. Some did not even do that. They simply stood there. No money, words or gesture. Just presence, and an expectation that something would begin without them having to start it.
-
And there were people who asked for advice. “What would you recommend?” “How many tokens are enough for an afternoon?” Not because they could not decide, but because they genuinely sought guidance from the person facing them.
What struck me was not the first group — they came prepared. Nor the third — asking for advice is reasonable.
It was the second group that struck me.
They had gone through the entire queue without formulating a request. They had outsourced the decision before even understanding the question.
That moment at the counter — someone handing over money without knowing what they actually want from the experience ahead — now plays out on a huge scale, millions of times a day, in chat windows and prompt boxes around the world.
Most conversations about AI concern speed.
Question → Answer → Done
That story is true — but incomplete.
Because if you start without a clear question, speed does not help. It accelerates confusion. You do not gain insight. You get fluent answers to unformed thoughts. Answers that sound good because the system is very good at sounding good.
Speed is not the problem. What disappears because of it is.
Something happens when a conversation pushes back. When someone says: “Wait — what exactly do you mean?” “Is that really what you are trying to solve?”
That friction — that slight resistance — is where thinking happens. Friction is not inefficiency. It is where a vague intention becomes a real question. It is uncomfortable. It slows things down. And it is precisely where many people’s real question lies — beneath the question they typed first.
The central question
What are the people building these tools actually optimising for?
Adoption rewards speed. Users per month. Searches per day. Tasks completed. A tool that answers immediately looks clever. A tool that asks three questions first feels slow — even if it leads to a better result.
The incentive is clear: answer quickly, solve quickly, feel useful. Even if the most useful thing you could do is say: “I am not sure you have thought this through yet.”
It was about how quickly people hand over the shape of a decision. How easily we trade clarity for convenience.
AI is now that someone. On the other side of the counter. Ready to tell you what to take.
The question is whether you came in with an idea of what you wanted — or have just let the tool decide for you, and called it productivity.
See. Do. Own.