You can try this!
check if you have a character, (a colon or space or both) that is followed by "
and by a \w(character or number).
>>> s
'{"name": "John", "company": "AB"C corp", "id": "12g: "123 12-12"}'
>>> re.sub('\w[: ]*"\w','',s)
'{"name": "John", "company": "A corp", "id": "1223 12-12"}'