Hello @Santosh Singh thanks for reaching out.
When you use fuzzy search, when the request is processed each word that it matches becomes a token, so for example searching on something like John*~* Smith*~ could translate into many different tokens such as Johan, Jahn, Smit, etc just to provide a few examples. AI Search won't weight the original token (exact match) of "John Smith" higher than other matched tokens, this seems to be the same case with* 道南, as fuzzy search might be creating many other tokens that come first.
In order to guarantee that the exact match takes priority, you could modify the query to look like this: (道南)^2 OR (道南~1) that way exact matches are boosted 2x but also including fuzzy search responses.
If you have any other questions please let me know.