I don't think this is true, do notation (see https://www.haskell.org/onlinereport/haskell2010/haskellch3....) specifically depends on the Monad typeclass and exists as syntactic sugar for monadic operators:

> A do expression provides a more conventional syntax for monadic programming