Gamma-ray bursts (GRBs) were first detected thanks to their prompt emission, which was the only information available for decades. In 2010, while the high-energy prompt emission remains the main tool for the detection and the first localization of GRB sources, our understanding of this crucial phase of GRBs has made great progress. We discuss some recent advances in this field, like the occasional detection of the prompt emission at all wavelengths, from optical to GeV; the existence of sub-luminous GRBs; the attempts to standardize GRBs; and the possible detection of polarization in two very bright GRBs. Despite these advances, tantalizing observational and theoretical challenges still exist, concerning the detection of the faintest GRBs, the panchromatic observation of GRBs from their very beginning, the origin of the prompt emission, or the understanding of the physics at work during this phase. Significant progress on this last topic is expected with SVOM thanks to the observation of dozens of GRBs from optical to MeV during the burst itself, and the measure of the redshift for the majority of them. SVOM will also change our view of the prompt GRB phase in another way. Within a few years, the sensitivity of sky surveys at optical and radio frequencies, and outside the electromagnetic domain in gravitational waves or neutrinos, will allow them to detect several new types of transient signals, and SVOM will be uniquely suited to identify which of these transients are associated with GRBs. This radically novel look at GRBs may elucidate the complex physics producing these bright flashes.