Question Regarding User typos when querying a Postgres DB using openAI

Hi All,
I’m working on application that connects openAI to retrieve data from Postgres DB based on the user’s question (in human language).
My main issue is that the user can sometime have typos in the questions.
For example, If I have a column in the data base called “retail_stores_name” and it has a various fields in it like Palisade mall, Hudson Plaza, President Square etc. The user can write something like (assuming there is another column called traffic which represents the amount of users visited in a particular date at each location)
" Give me the top 5 days with the most traffic at pallasaide sq"
Notice that there is a typo in the location as well as the wrong name (should be mall, not square).
I notice that OpenAI wasn’t able to deal with this prompt. Could someone please advise on that?

You can also use the intelligence of the AI - tell it about the need for exact matches, and the error-prone input of a user at a keyboard. You can either do a “did you mean” detection on the unreliable query field or the lack of results, or simply have the AI not send without clarification if it is a questionable input.