The formation, fuelling and feedback of active galactic nuclei (AGN) are fundamental ingredients in our picture of galaxy evolution, and yet many mysteries remain regarding the physical processes that govern them. For example, whilst there are numerous theoretical mechanisms that can fuel an AGN, ranging from bars, to disk instabilities and cosmic accretion, an observationally proven model has yet to emerge. In this talk, I will present a multi-wavelength, observational perspective focusing on the connection between AGN and galaxy mergers. I will show unequivocal proof that mergers can trigger both optically and mid-IR selected AGN, and that these AGN are triggered preferentially in the late stages of the merger. On the other hand, I will argue that low accretion rate AGN that are selected in the radio regime (the so called “radio mode” or radiatively inefficient” AGN) are NOT linked to mergers. Instead, two secular processes are required to explain the observations of low excitation radio galaxies: one from an internal stellar source, such as winds from evolved stars, and one from a larger scale source, such as a hot gas halo. These results support the picture of two fundamental modes of AGN fuelling. I will also show results from an on-going search for binary black holes within our late-stage merger sample. Finally, I will discuss the levels star formation in AGN hosts – another controversial topic in recent years. Using artificial neural networks to predict the IR luminosity of over 330,000 galaxies in SDSS, we have unprecedented statistics to tackle this question and can now quantify how star formation rates in AGN host galaxies differ (or not) from ‘normal’ hosts.