Based on n sample size (if n is greater than or equal to 30 it is normally distributed -Central Limit Theorem) how do you know if one mean inference is valid? And also how do you know based on distribution if t or z procedures should be used?

Please explain step by step I am so confused