When a paper is presented at a conference, it is generally not complete. The general practice is to present a preliminary analysis at the conference and reserve the detailed analysis for a full paper. The purpose of presenting your research at a conference is to inform people about your study and get feedback from them. So once you receive the inputs, you might want to incorporate them in your paper. Thus, the paper that you will submit to the journal will definitely be an improved and a more detailed version of the one you presented at the conference.
Regarding the extent and nature of revisions, it is advisable to change the title so that people do not confuse the conference paper for the journal paper or vice versa. Additionally, make all the enhancements based on the feedback of the conference review process and the discussion at the conference. Since conference papers generally state the preliminary results, include any new results in your journal paper. The percentage of revision required may be field specific and journal specific, but a minimum of 30 per cent revision is a must in most cases. Carefully go through the submission policy for previously published conference papers that are usually provided under author guidelines and make sure your paper follows all the guidelines.
Thanks Osama Rahil for your feedback. They are some conferences that do not thoroughly check papers presented in their conference before publishing. Not minding if the author changes the title or not. In this case, in what category would the article fall?
It counts as a journal article which is more important. It might also be mentioned as a conference paper as well. The two versions are likely to different any way, as has been said.
It is also worth keeping the titles slightly different so noone thinks you're putting identical work in two categories. If the two versions are actually identical, I'd just list it as a journal article.