Using botpress integrated with make and elevenlabs, you can send an audio as if it were recorded on the spot, to appear as humanized as possible?