This question is about the definition of conserved quantities integrable systems.
Using Algebraic Bethe ansatz,a family of commuting operators $F(\lambda)$ can be contructed by taking a partial trace (over the auxiliary space) of the monodromy matrix $T(\lambda)$
$$F(\lambda) = tr(T(\lambda)),$$
$$[F(\lambda),F(\mu)] = 0,$$
as stated in Faddeev's paper.
Furthermore, they should be "functionally" independent; this concept is clear in classical integrable systems but I can't find a good definition for the quantum case.
In principle, one can construct an extensive number of conserved quantities for a finite dimensional Hilbert space by considering projectors onto each eigenspace $P_{n} = |n\rangle \langle n|$
$$[P_n,H]=0,$$
this argument is independent of $H$.Therefore, having an extensive number of conserved quantities should not be a good definition for integrability.
These conserved quantities should probably be local (i.e. sum of few-body terms), but it is not clear if the trace of the monodromy matrix always gives such local operators.