Extracting accurate photometry (and astrometry) from images taken with adaptive optics assisted instruments is particularly challenging. Current post-processing tools are not prepared to achieve high accuracy from AO data, especially in limiting cases of crowded fields and marginally resolved sources. We quantify the limitations of these tools with synthetic images, and present a proof-of-concept study showing the potential of using reconstructed PSFs from the (GL)AO system telemetry to increase the measured photometric accuracy. We show that the photometric accuracy is significantly improved with a good PSF reconstruction in considerably crowded regions. We demonstrate the need for a dedicated post-processing tool that incorporates available information about the PSF, as well as the ability to adjust to the spatial variations of the PSF characteristic of AO data.