Occam's Razor

Would the following arrangement violate Occam's Razor:

- A single methodology that initializes with the assumption the 'answer'
will be complex not simple (maybe assumes some sort of maximum or
infinite-in-the-limit complexity), but progressively approximates to the
simplest possible explanation?

The way I see it, this setup is legitimate in Occam terms, if the
legitimacy of an arbitrary occam-correct explanation is unaffected by
the origin of that explanation. For example, it could come by a random
guess, or a creative process....or indeed some sort of 'guiding' method.

Opinions appreciated.