What is the best way to separate text from ai output?

Is there a simpler way to have open ai output a results and pull certain things from it?

An issue I ran into was that a team member want to format the results that open ai outputted by there is no control over individual words in the output. Which then required me to run the API multiple time just to have something like the body and text be formatted differently from each other.

Is there an easier and more efficient way?