There is usually one person sitting with their arms folded. They did not ask to be here, they have a view about what building with LEGO has to do with their job, and they are being polite about it. That is a completely reasonable starting position, and it is worth saying plainly that most people who write about the method afterwards describe arriving skeptical and leaving surprised.
So rather than sell the feeling, here is the sequence.
Everyone builds before anyone talks
A question goes to the room, and the first thing that happens is not discussion. Everybody builds something in response to it, at the same time, individually. Nobody is watching anyone else's hands, nobody is waiting for a turn, and crucially nobody has heard anyone else's answer yet.
That last part is doing most of the work. In a normal meeting the first person to speak sets the frame, and everything after that is a reaction to it. When the room builds first, you get genuinely independent positions instead of variations on whatever was said at the start.
Every model gets heard, without exception
Then each person explains what they built and what it represents. Not a summary from the facilitator, not a show of hands — each person, in turn, describing their own thing.
Two useful things happen here. The obvious one is that the quiet people speak, because the structure hands them the floor rather than requiring them to take it. The less obvious one is that people say more than they would have said in open discussion. Talking about a model puts a small amount of distance between the person and the point — it is the model that has the problem in it, not you — and that distance makes it noticeably easier to name something difficult.
The group builds the thing it has to decide
Individual models then get worked into a shared one. This is where the actual argument happens, and it is a much better argument than the usual kind, because it is concrete. Instead of "I think we are misaligned on priorities," someone is physically moving a piece and saying where it should sit and why.
The output is not a photo of some LEGO. The output is an agreed picture of the situation that everyone in the room helped construct, plus the decisions that fall out of it, written down as things people have committed to.
Why the hands matter
The method leans on a straightforward idea: thinking with your hands surfaces things that talking does not. One first-hand account put it as unlocking knowledge stored somewhere you cannot easily reach by discussion alone. You do not need to accept any grand claim about that to see the practical version — people who cannot immediately say what is wrong can very often build it, and then explain what they built.
It also removes the thing that wrecks most sessions. When everyone has built and everyone has spoken, a dominant voice has far less room to run, because the airtime was allocated by the structure before anyone opened their mouth.
When it is not the right tool
It is worth being equally clear about the other side. If the topic is simple, this is more process than the problem deserves. If there is one correct answer, an analytical tool will find it faster. If the decision has already been made and the session is meant to produce agreement with it, this is actively the wrong choice, because the method generates answers rather than delivering them.
If you want a straight read on whether your situation fits, the fit assessment is ten questions and will tell you no when no is the right answer. If it does fit, the session formats are here.