We investigate the impact of combining Gaia astrometry from space with precise, high cadence OGLE photometry from the ground. For the archival event OGLE3-ULENS-PAR-02, which is likely a black hole, we simulate a realistic astrometric time-series of Gaia measurements and combine it with the real photometric data collected by the OGLE project. We predict that at the end of the nominal 5 years of the Gaia mission, for the events brighter than $Gapprox15.5$ mag at the baseline, caused by objects heavier than 10 $M_{odot}$, it will be possible to unambiguously derive masses of the lenses, with accuracy between a few to 15 per cent. We find that fainter events ($G<17.5$) can still have their lens masses determined, provided that they are heavier than 30 $M_{odot}$. We estimate that the rate of astrometric microlensing events caused by the stellar-origin black holes is $approx 4 times 10^{-7} , rm yr^{-1}$, which implies, that after 5 years of Gaia operation and $approx 5 times 10^6$ bright sources in Gaia, it will be possible to identify few such events in the Gaia final catalogues.