A significant fraction of the bolometric luminosity of NLS1 is in their steep soft X-ray spectral component. Recent Chandra and XMM data allowed to resolve soft X-ray emission/absorption lines but at the same time showed that bulk of the soft X-ray emission is just a continuum, most probably a result of Comptonization of the accretion disk emission. However, it opens a question where this Comptonizing region forms and what is its origin. I would like to discuss in detail the two most plausible scenarios: accretion disk outflow and a hot disk skin. The first scenario is supported by large shifts of the warm absorber lines in some objects (Pounds et al. 2003) but these results strongly relay on proper identification of the most intense lines. Our models of the warm absorber indicate that the Ly_alpha lines from CNO elements not always dominate. The second scenario is supported by the analysis of the radiation pressure dominated disk. In such a disk considerable fraction of energy must dissipate in the upper disk layers possibly leading to formation of a hot disk skin (e.g. Czerny et al. 2003, Turner 2004). I would like to discuss the possibilities to differentiate between the two options.