Finding high redshift AGN is important for constraining cosmic evolution and reionisation. In recent years it has become clear that the numbers of AGN at redshifts above 3 has been underestimated by a factor of at least 30%. Our group has used various methods, including color, x-ray emission and machine learning recipes to identify AGN at z>4. Based on the results of our campaigns at z~4 and ~5, the derived space density of AGN is higher than previous works and this leads to an important contribution of the population to the IGM ionization. Part of these results are based on the QUBRICS survey (QUasars as BRIght beacons for Cosmology in the Southern hemisphere) that has been active since 2019, aiming to identify bright (i_mag<18.0) and high redshift (z>2.5) QSOs in the Southern Hemisphere. Apart from improving the estimate of the luminosity function, this new database has enabled multiple scientific applications that will be briefly highlighted during the talk.