At Verbatim, Baby!, we see the need for better representation in African languages every day. Existing ASR models are very effective for unaccented English audio, but not so much for South Africa's ten other official languages.