The latter case is what compression is for: to even-out variations in volume.
The LevelSpeech2.ny plugin is good for podcasts, it’s a compressor-limiter combo.
[ The optimum compressor settings will be different for speech and music: i.e. they should be processed separately ].