How much the AGN contributes to the far-infrared emission of a
galaxy has been a topic of contention for many years, recently
re-ignited due to the vast amounts of excellent quality far-IR data from the Herschel space observatory. Disentangling the contribution of the AGN and star-forming components in a galaxy's energy budget directly affects the estimation of star-formation rates and if we are to accurately measure the star-formation rate density as a function of redshift then we must determine the form of the intrinsic AGN SED. In addition, understanding the star-forming properties of QSO host galaxies will be a major step towards understanding galaxy evolution. Until now, results and views on this topic have varied greatly according to the samples examined, the models used and selection biases which have often been ignored. Here, I will talk about our recent results regarding the emission of AGN up to submm wavelengths and present the intrinsic AGN SED we derive with a sample of powerful QSOs. Using this SED, I will show that in many cases SFRs have been overestimated, leading to over-interpreted correlations between the SFR and AGN power.