Why not pass the inverse of the world matrix as a variable to your shader?
And normally you would be multiplying the point by the world, view, and projection matrices to transform from object space to world space. Is that what you mean to do?
EDIT: If you still intend to invert the matrix, it would get very slow to invert the matrix for every vertex, every frame. The best way to go about this is to pass the preinverted matrix to the shader once.