Expected $R_\mathrm{AA}$ for the measurement of fully reconstructed $\mathrm{B_s^0}$ mesons with the ITS2 and ITS3 detector upgrades compared to TAMU models. Four transverse momentum bins into reach with ITS3 setup, two for ITS2 setup.

 - Analysis performed exploiting the BDT of XGBoost.
 - MC sample: LHC19h1b (signal and combinatorial background source) and LHC19h1a (true Ds+pion background source). Full simulation with ITS2 geometry produced with HIJING underlying event + pythia6 events with ccbar and bbbar pairs (signal)
 - Improver task applied to smear the tracks and update the track parameters with the ITS3 resolution.
 - The statistical uncertainty assumes that the Pb-Pb one is dominant
 - The systematic uncertainties are estimated directly for the $\mathrm{B_s^0}$ mesons where possible, or coming from an educated guess using the $\mathrm{D_s^+}$ study with XGBoost in Pb-Pb 2018.
 - Compared to TAMU predictions for strange and non-strange B mesons received from private communications with the authors. Model corresponds to [PLB 735 (2014) 445].