Let's set some context: your eyes typically "converge" when looking at an object. The closer the object is (like your nose), the more pronounced your eyes will rotate inward to view it. The further away, the more your eyes will point parallel to each other.
The famous Magic Eye book uses this behavior to create optical illusions. The trick is that you want to relax your eyes and look "through" the book to see the hidden picture. Here's three ways your eyes behave:
What you're rewarded with is a 3D image that emerges from what is seemingly chaos.
How does this relate to product development?
It doesn't take much of a leap to realize: your eyes are your team, and the picture is your product. Letting the team go in any direction they want can be helpful, but only to a point. Exploration is encouraged, "do I see a giraffe?" "what do you see over there?" but after a while, results need to emerge.
If anything, our new AI-shaped world has exacerbated the over-exploration problem, but this should be unpacked with a different post.
Even simply having your eyes look as they always do (aligned vergence) can be useful, "we've always done it this way!" your team may tell you. They're not wrong, but if this was about building something that already existed, what's the point of this exercise in the first place? What you want is off in the distance and you all need to work together to get there.
As for cross-eyed vergence, well... I think we've all seen how those teams operate.
With just the right amount of direction—and maybe even a bit of indirection—something will start to take shape as you look through this mess of a pattern. If one eye starts to veer, the image is lost. Ironically, when your eyes are pointed in the same direction, it's called "teaming."
The question you have to ask yourself is: is your team diverging or converging? Are you on the path to see that 3D image or is everyone staring at a wall headed in completely different directions? Only when you get convergence will you get a cohesive product out the other side.