This is mostly a question to be honest... I've been going through the docs learning about RTK and everything is pretty much clear with nice examples. You create a query or mutation, and these can be triggered using the autogenerated hooks in components like so:
https://redux.js.org/tutorials/essentials/part-7-rtk-query-basics#using-query-hooks-in-components
const { data: posts = [], isLoading, isSuccess, isError, error } = useGetPostsQuery()
Later, at RTK Query advanced patterns we learn that we can still use createEntityAdapter to manage normalized data:
https://redux.js.org/tutorials/essentials/part-8-rtk-query-advanced#transforming-responses
and we can use
export const { selectAll: selectAllUsers, selectById: selectUserById } = usersAdapter.getSelectors(selectUsersData)
to start using selectors of our entity.

What is not clear to me is how we use the normalized data in our components? Do I convert the above to:

const { isLoading, isSuccess, isError, error } = useGetPostsQuery(); const posts = useSelector(selectAllPosts)
I guess the question is, do I stop using the data exposed by the generated query hook and start calling the selectors generated by the entity adapter? Or is there some kind of integration that I am missing which updates the data returned from the generated hook? I think I am mixing the concept of querying with normalization which are separate and I am guessing it makes sense to be kept this way.

Thanks for making our life easier :) Migrating from older redux version to RTK and we can already see the benefits and the boilerplate cutdown.