A transcript is the accessible version and the searchable one
The same document solves two different problems at once, which is why it is not optional.
Audio is unusable for anyone who cannot hear it. Not degraded, not inconvenient: unavailable. There is no partial access to sound the way there is partial access to a page with small text. If the explanation exists only as audio, then it has been made for some people and withheld from others, and no amount of quality in the sound alters that. A written equivalent is the mechanism by which the content becomes available to everybody, and it is an obligation rather than a courtesy.
The same document turns out to solve a completely separate problem for people who can hear perfectly well. Sound cannot be searched. A listener who remembers that something useful was said, but not where, has no way to find it in audio and every way to find it in text. Scanning, skimming, quoting, checking a claim, returning to a passage a fortnight later: all of these are ordinary uses of an explanation, and none of them are possible until the words exist in a written form.
That double duty is the argument for treating the written companion as part of the work rather than as an afterthought produced when time allows. It also raises the standard: a written version that is unreadable serves the searching use poorly and the access use worse. Structure it, mark where the topics change, and make sure that anyone who reads it instead of listening comes away with the same understanding, which is the actual test of whether the equivalent is equivalent.