Let $\rho_1$ and $\rho_2$ be two representations of $G$, $V$ and $W$ be the correspondence $FG$-modules, respectively.
Recall the tensor product $\rho_1\otimes \rho_2$ of $\rho_1$ and $\rho_2$ is defined by $\rho_1\otimes \rho_2:G\to GL(V\otimes_F W)$, $(\rho_1\otimes \rho_2)(g)(v\otimes w) =(\rho_1(g)\otimes \rho_2(g))(v\otimes w) =(\rho_1(g)(v),\rho_2(g)(w))$.
On the other hand, suppose that the $FG$-module $V\otimes_{FG} W$ afford the representation $\rho_3$.
My Question: Is $\rho_3\sim \rho_1\otimes \rho_2$?
$~~~~~~~~~~~~~~$ $FG$-module $V$ $~~~~~~~~~~~~~~$ $\to$ $FG$-module $V\otimes_{FG} W$ $\leftarrow$ $~~~~~~~~~~~~$ $FG$-module $W$
$~~~~~~~~~~~~~~~~~~~~~~~~~~$ $\updownarrow$ $~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~$ $\updownarrow$? $~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~$ $\updownarrow$
representation $\rho_1:G\to GL(V)$ $\to$ $~~~~~~~~~~$ $\rho_1\otimes \rho_2$ $~~~~~~~~~~~~~$ $\leftarrow$ representation $\rho_2:G\to GL(W)$
The tensor product of the $FG$ modules is the tensor product over the field $F$. This might seem strange, as you were expecting a tensor product of bimodules and hence wrote a tensor product over $FG$.
But consider two algebras $A$ and $B$ with modules $V$ and $W$. Then $V \otimes_F W$ is an $A \otimes_F B$ module in the evident way. We are going to take $A = B = F G$. But how does $V \otimes_F W$ become an $A$ module when it should be an $A \otimes A$ module? There is an algebra map $A \to A \otimes A$ extending $g \mapsto g \otimes g$, and one composes this with the representation of $A \otimes A$ on $V \otimes W$, ending up with the representation $\rho_1 \otimes \rho_2$ as you have described it.