Résumé : The detection of gravitational waves (GW) with an electromagnetic counterpart enabled the first Hubble Constant $H_0$ measurement through the standard siren method. Current constraints suggest that $\sim 20-80\%$ of LIGO/Virgo/KAGRA (LVK) Binary Black Hole (BBH) mergers occur in Active Galactic Nuclei (AGN) disks. The claim for a possible association of several BBH mergers with flaring AGNs suggests that cosmological analyses using BBH and AGNs might be promising. We explore standard siren analyses through a method that takes into account the presence of background flaring AGNs, without requiring a unique host galaxy identification, and apply it to realistic GW simulations. Depending on the fraction of LVK BBHs that induce flares, we expect to constrain $H_0$ at the $\sim 3.5-7\%$ ($\sim 2.5-5\%$) precision with $\sim 2$ years or $\sim 160$ events ($\sim 1$ year or $500$ events) of LVK at design (A+) sensitivity, assuming that systematic BBH follow-up searches are performed. Assuming a more restrictive $\Omega_{\rm m}$ prior and that at least $20\%$ of BBHs produces detectable flares, we may reach a $3\%$ ($2\%$) precision in $H_0$ after 2 (1) year of LVK at design (A+) sensitivity. We also show that a $\sim 5-10\%$ precision is possible with complete AGN catalogs and 1 year of LVK run, without the need of time-critical follow-up observations.