Form factors are a crucial theory input in order to extract |V_cb| from B → D^(*)ℓν decays, to calculate the Standard Model prediction for R(D^(*)) and to assess the impact of new physics. In this context, the dispersive matrix approach, a first-principle calculation of the form factors, using no experimental data but rather only lattice QCD results as input, was recently applied to B → D^(*)ℓν. It predicts (within the Standard Model) a much milder tension with the R(D*) measurements than the other form factor approaches, while at the same time giving a value of |V_cb| compatible with the inclusive value. However, this comes at the expense of creating tensions with differential B → D*ℓν distributions (with light leptons). In this article, we explore the implications of using the dispersive matrix method form factors, in light of the recent Belle (II) measurements of the longitudinal polarization fraction of the D* in B → D*ℓν with light leptons, F^ℓ_L, and the forward-backward asymmetry, A^ℓ_FB. We find that the dispersive matrix approach predicts a Standard Model value of F^ℓ_L that is in significant tension with these measurements, while mild deviations in A^ℓ_FB appear. Furthermore, F^ℓ_L is very insensitive to new physics such that the latter cannot account for the tension between dispersive matrix predictions and its measurement. While this tension can be resolved by deforming the original dispersive matrix form factor shapes within a global fit, a tension in R(D*) reemerges. As this tension is milder than for the other form factors, it can be explained by new physics not only in the tau lepton channel but also in the light lepton modes.