Suppose I have two matrices
$$A = \begin{bmatrix} 0 & 1 \\ a_2 & a_1\end{bmatrix}$$ $$B = \begin{bmatrix} 0 & 1 \end{bmatrix}$$
Then $AB^T = \begin{bmatrix} 1 \\ a_1\end{bmatrix}$, the last term is $a_1$
Now I expand matrix $A$ and $B$ a little bit to get
$$A = \begin{bmatrix} 0 & 1 & 0 \\ 0 & 0 & 1 \\ a_3 & a_2 & a_1\end{bmatrix}$$ $$B = \begin{bmatrix} 0 & 0 & 1 \end{bmatrix}$$
Then I find $AB^T = \begin{bmatrix} 0 \\ 1 \\ a_1\end{bmatrix}$, the last term is $a_1$, furthermore $A^2B^T = \begin{bmatrix} 1 \\ a_1 \\ a_1^2 + a_2 \end{bmatrix}$, the last term is $a_1^2 + a_2$, neato!
Finally let's define our matrix to be: $$A = \begin{bmatrix} 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 1 \\ a_4 & a_3 & a_2 & a_1\end{bmatrix}$$ $$B = \begin{bmatrix} 0 & 0 & 0 & 1 \end{bmatrix}$$
As before, I find $AB^T = \begin{bmatrix} 0 \\ 0 \\ 1 \\ a_1\end{bmatrix}$, the last term is $a_1$, furthermore $A^2B^T = \begin{bmatrix} 0 \\ 1 \\ a_1 \\ a_1^2 + a_2 \end{bmatrix}$, the last term is $a_1^2 + a_2$
Curiously, I can compute $A^3B^T$ and obtain $\begin{bmatrix} 1 \\ a_1 \\ a_1^2 + a_2 \\ a_3 + a_2a_1 + a_1(a_1^2+a_2) \end{bmatrix}$
Now suppose A and B are both N dimensional matrices.
Is there a closed form solution for the n by nth value of the product $A^{n-1}B^T$?
Note: $A$ and $B$ pair are what known as the canonical controllable form