At the moment, for UV/Vis electronic spectra Avogadro just applies a Gaussian broadening of a specified FWHM (converting that to a \sigma for the Gaussian), using the oscillator strengths as the intensity of the original peaks.
I don’t have a screenshot for UV/Vis spectra, but I recently posted an example IR with the stick spectra indicated Feature Wishlist for a Spectra Dialog - #5 by ghutchis
There can be a few reasons for different types of broadening in spectra, but at the moment, we’re mainly aiming for the “looks closer to experiment because of various broadening effects”
Does that help?