As for question 2
Edit: I haven't nailed down all the details, but if $G$ is the lift of $\hom(X, -): M \to Set$ provided by a co-$T$-algebra structure on $X$ internal to $M$, then provided that the canonical natural map
$$\hom_M(A, B) \had responded to \hom_M(GA, GB)$$
is an algebra map (as I think it will bequestion 2 earlier, but these are the details I haven't fully checked), then there will in fact be a canonical isomorphism $[X, -] \cong G$ where $[-, -]$ is the internal hom in $M$ (here $[X, -]$ provides the "usual" factorization). Provided am now editing that $G$ is thus enriched, the idea response out as it is first to produce a canonical arrow $[X, -] \to G$; superseded by the enriched Yoneda lemma, this corresponds to an arrow $I \to G(X)$ where $I = F(1)$ is the monoidal unit of $M$; this in turn corresponds to the canonical arrow
$$1 \to \hom_M(X, X) \cong UGX$$
that names the identity on $X$. Then, applying $U$ to $[X, -] \to G$, we get the isomorphism $id: \hom(X, -) \to \hom(X, -)$, but since a monadic functor $U$ reflects isomorphisms, we see that $[X, -] \to G$ already had to be an isomorphism. In other words, up to isomorphism $G$ coincides with the "usual" lift $[X, -]$.
Continuing this train of thoughtMartin's second comment below, suppose $\{\theta_i: X \to \sum_{ar(\theta_i)} X\}$ and $\{\eta_i: X \to \sum_{ar(\eta_i)} X\}$ are two coalgebra structures on $X$ which lift $\hom(X, -)$ to $M$-algebra functors $[X, -], [X, -]': M \to M$ respectively. By the above, we know that $[X, -]$ and $[X, -]'$ are isomorphic; more exactly that there is an algebra-valued isomorphism $[X, -] \to [X, -]'$ lying over makes the identity $\hom(X, -) \to \hom(X, -)$ (and there is just one such transformation since $U$ is faithful)situation quite clear. But this says precisely that the identity on $X$ is a coalgebra map from the one coalgebra structure $\{\theta_i\}$ on $X$ to the other $\{\eta_i\}$, which is just another way of saying Apologies for the two coalgebra structures on $X$ coincidenoise.

