When and where would it evolve, and why?
To be "RF", we must mean frequencies far above the cutoff due to ionic diffusion. This isn't a neural thing, this isn't a muscle contraction thing. Eels are scrunchy batteries, not spark gaps.
It has to be pretty fundamental, an early-ish discovery; most mechanisms I can think of involve some kind of mechanical interaction, piezoelectricity of a specific crystal for example, combined with some highly nonlinear process for harmonic generation (e.g. spark breakdown, but less destructive and more consistent).
I mean, that's not even that far out, hydroxyapatite (the mineral of vertebrate bones) is piezoelectric, at least a little bit. But it would have to be applied in some way to get a transient into it, and have a spark gap or avalanche breakdown effect around it, with enough electrical insulation not to spoil the effect while also maintaining adequate structure and perfusion to, you know, be a useful otherwise-ordinary part of the organism.
And then there has to be a simultaneous mutation that effects some kind of receiver.
And these structures will be small, like, multicellular organisms, whether invertebrate or otherwise. So any relevant frequencies will be very high, or very close in.
And the most dramatic evolution all took place underwater, where there was plenty of optical transparency, but RF is wholly shorted out.
But even this is very unrealistic, very mechanistic, conventional minded; it's probably more likely that protein structures might evolve to create organic semiconductors already, and perhaps diodes (of various types, including the potential for negative resistance) and transistors could be created. Powered by muscle contractions, piezoelectricity, proton pumps (or other ions, but protons are fastest of course), etc., perhaps signals could be created idiomatically -- by random chance, and with similar junctions between individuals, perhaps enough reciprocity could be present that the transceiver problem is solved.
But this still assumes that there is an environment conducive to RF transmission (so, basically terrestrial only), that such structures evolve despite (or at best in parallel with) competing solutions that are easier to develop; like small-molecule emission and reception (smell!), but also optics have the advantage that quantum effects like chemiluminescence can be discovered by random chance, and, light sensors have been a staple of most/all epipelagic and terrestrial life forms, probably since before predation evolved, even?
I guess there's kind of an interesting angle (and, equally well I suppose: reason to be skeptical of any such proposed story's potential accuracy, and space to be creative with what kinds of ideas might be possible) to "alternative history" evolution, that kind of parallels a lot of mythologies, like: there once was an age of fantastic beasts of all sorts, but things get more familiar, less miraculous, more banal, more standardized, as you approach the present. Why don't you see elves and goblins anymore? Same reason you don't see trilobites, or Dickensonia, anymore: they all "moved into the west". Why don't they evolve again now? Simple: there's no pressure to; there's no mechanism to make such radical changes, under pressure of everything else in the environment they have to fit in with. The basic vertebrate body plan hasn't changed in over 300 million years or so, and it's just so foundational there's essentially no way for mutations to occur in those lowest-level genes without succumbing to miscarriage (or the oviparous equivalent). In fact it's far easier to severely deform those structures to new uses, than to redo the fundamentals of the body plan (see horse legs and hooves for example, the wrist and hand being largely resorbed or fused in favor of the middle-finger bones for weight bearing).
Well, along such a timeline, I would suppose such a special structure would've had to evolve quite early, but there explicitly weren't the conditions necessary to permit its evolutionary testing/exploration.
Also worth mentioning, electron/ion exchange in bacterial mats, whether through biofilaments, or inorganic (ferrous?) or hybrid structures; but this is basically DC, and I'm not aware of any that are particularly conductive, nowhere what you'd need to make a resonant structure as such. Alternately, evolution would have a field (ha!) day with resonant cavities and lenses and antennas, that would be quite cool to see across species -- but again, without a way to harness microwaves, nor a medium to carry them, such ideas will remain confined to ones' imagination.
Tim