I want to expand the axis of a ndarray to a diagonal matrix in the fashion of diagflat

E.g.

In [ ]: import numpy as np

In [ ]: example = np.random.random((200, 5))
In [ ]: example.shape
Out[ ]: (200, 5)

What I am looking for, is something like:

In [ ]: np.diagflat(example, axis=-1).shape
Out[ ]: (200, 5, 5)

diagflat has however axis argument. My idea was the to simply insert a new axis in for example and multiply it with a unit matrix.

In [ ]: Id = np.eye(example.shape[-1])
In [ ]: (example[..., np.newaxis] @ Id).shape
ValueError: shapes (200,5,1) and (5,5) not aligned: 1 (dim 2) != 5 (dim 0)

This raises however an error, apparently broadcasting is not applied for matrix multiplications. Is there an elegant solution, or do I have to create and fill the array by hand?