As I stated in my answer to the linked question, the evolutionary prerequisites for radio communication is in a species that is able to precipitate a variety of metals in a variety of forms and an environment high in metals.
Initially, a precursor species would evolve to use metal to enhance its neural transmission rates as electrical transmission is vastly faster than human nerves' sodium-gate depolarization system, in itself a highly advantageous strategy in evolutionary terms.
It is likely that creatures using metal as a nerve conduction rate booster would have found that unshielded nerves would cause radiation detectable not only within a creature's own body, but in other creatures too. As faster nerve conduction is too great an advantage to give up, shielding would have evolved, quite possibly by running nerves through the centres of metal bones, or perhaps by sheathing the individual neurons in metal.
However, the possibilities of transmission and detection of EM radiation means that not all metal nerves would have evolved to be completely shielded, some could be partially shielded and be used to detect EM radiation.
So, we have evolved creatures that emit RF energy as a by-product of their neural activity. From there, once shielding has evolved to reduce cross-talk between nerve fibres, detection of RF leakage requires more sensitive receiver organs. Along with this, any deliberately unshielded neurons would emit RF energy detectable at greater range.
As there is almost always an advantage in being able to communicate at longer distances, the evolution of a stacked pile of depolarising cells (as occurs in electric eels) allows higher transmitter voltages, and hence higher power and range.
We then get to the point of bandwidth. EM radiation emission will most likely begin at lower radio frequencies, but it is entirely possible that mechanisms could evolve to increase the frequency of emitted radiation. Since a system of this type could have practically each neuron driving an EM transmitter of a different frequency, high bandwidth can be achieved by rapid changes in signal amplitude and frequency that is allowed by using high-frequency EM radiation, and also by multiplexing - using many frequencies simultaneously. This could ultimately allow an evolved bandwidth many times greater than our own Wi-Fi communication, which could also be somewhat directional. Another argument for higher radio frequencies and microwaves is that smaller antennas are required.
Since all this bandwidth is relatively easily achieved in evolutionary terms - simply by duplicating the relevant organs - there is no reason why the beings would not evolve to make use of this bandwidth. Since the highest intelligences of species on earth are found in those creatures with an active social life (and this ability makes for a great social life), the evolution of intelligence is pretty much a given.
Considering that a sentient, tool-using species that can communicate via RF at what are probably high bandwidths, it is unlikely that humans could easily develop an interpreter for this alien language, especially given that it would most likely be multiplexed, and both frequency- and amplitude-modulated, as well as rapid and idiosyncratic rather than following any simple grammar as in human-manufactured RF communication. It is far more likely that these creatures - should it occur to them that audio is being used to communicate ideas, a not-unlikely proposition given their inherent ability to share processing - would learn to understand and communicate with humans using human language, given its likely lower bandwidth and complexity.
Of course, since we're talking about evolution, an evolutionary feature such as metal-enhanced neurology would have to occur at a very early point in the species' evolutionary history. This means that - thanks to evolutionary divergence - there would most likely be a great number of species on this world which emit RF energy to a greater or lesser extent.
We can anticipate that in the groups of creatures with unshielded neurology, the 'noisiness' of their neurons would be a beacon to the RF senses of predators, particularly those who have shielded neurons themselves, and thus have a lower background noise over which they can 'hear' their prey. From this, we can anticipate that these may be easy prey to such predators, and would hence be prone to adopting an r-strategy.
Other species would have evolved to communicate via RF to a greater or lesser extent; we can anticipate a wide variety of such creatures occupying multiple niches, though as the communication range of RF is such that it can be anticipated that many would be at least a bit smarter than a terrestrial-equivalent species mainly due to the greater opportunities for social interaction.
As to the environment, there is practically a necessity for more metals to be accessible. This does not preclude an oxygen atmosphere, but there may be levels of atmospheric dust containing heavy metals that would have toxic effects on humans not protected by respiratory filters or who eat the local life forms. This would make face mask filters highly advisable rather than essential, and we could have a human living for quite some time without one before they might start to experience symptoms of heavy metal poisoning.
A potential reason for humans to be interested in such a world is that with the biological precipitation of metals, mining metals would be an almost trivial exercise of picking up the carcasses of dead creatures, whether recently dead or fossilised. Some very interesting alloys are likely to have evolved, as is foam-metal which is both light and strong due to its internal voids.
