While the terms agonist and antagonist have rigorous meanings in pharmacology, we use the terms more loosely when naming built-in dose-response equations.
We use the term agonist to describe curves that go uphill, with a larger Y value as the X value gets larger.
We use the term antagonist to describe curves that go downhill, with a smaller Y value as the X value gets larger.
We use the term normalized to describe curves that go between Y=0.0 and Y=100.0. Even if you have normalized your data, you may not want to use the normalized equation. One reason not to use the normalized equation is to let Prism fit the top and/or bottom plateaus. Another reason is that with some drugs (a partial agonist) the top plateau won't be 100 (assuming you define 100 based on the response to a different drug).