An idea that probably has occurred to many is using reduplication for marking habituals. However, what about reduplicating the tense marker instead? Now, a different option appears: don't just reduplicate it, apply it twice instead.
What is the difference? Reduplication takes phonological matter and copies it, with some possible morphophonological rules applied. Applying it twice applies all morphological rules that are relevant, and then morphophonological rules. Thus, if there's two verb conjugations, and the first person singular past verb in the first conjugation looks like a second conjugation stem, the first person singular past habitual would consist of a first conjugation suffix followed by a second conjugation suffix.
Different persons may behave differently with regards to that, due to the first inflected form appearance possibly deciding which conjugation the second suffix takes. Of course, one could have more conjugations that interplay in complicated ways as well. Other morphophonology could of course also apply.
One could of course consider similar things for plurals - duplicate case suffixes or gender suffixes, and do so by rules that make them vary a bit at times. However, this creates a fun situation with regards to the nominative, a case that oftentimes is not marked by any explicit morpheme. Maybe the reduplication then defaults to reduplicating the root or some syllable of it - or uses a different case marker, e.g. the accusative, thus conflating the two in the plural (not an unusual thing in the languages of the world).