I'd probably use timeframes to delineate between different modes/pillars of play.
Combat - action declarations cover the next few seconds, action declarations and resolution happen in a pre-determined order and all participants take equivalent turns.
Active exploration - generally action declarations cover next few seconds to few minutes, depending on the hazards involved. The order of resolution of individual actions is generally more freeform, and often not all participants consume equivalent actions. Exact dialog between PCs and NPCs would fall in here.
Passive exploration/travel - Action declarations usually cover several hours, and are generally done by the entire group, with some participants taking "roles" (scout, rearguard, etc.) Explicit actions generally folded into the overall time frame. ("I cast some healing spells while we walk.") Discovering hazards moves back into one of the two previous modes. Longer-length social dynamics, which don't require exact dialogue, happens here. ("I haggle with the merchant for 10 minutes over the price of the armor." "I ask various bar patrons about what they saw for the next half hour.")
Downtime - Action declarations in the days to weeks time frame, primarily used to reframe the next section of the narrative. Crafting, healing, and domain management can all occur here if desired.