How does evacuation behaviour of crowds vary based on the level of perceived urgency? This question was investigated using two simulated experiments of crowd evacuations, collectively exploring three important aspects of evacuees’ decision-making: reaction time, exit choice and exit-choice adaptation. The analysis of reaction time observations identified a range of factors that determine the delay between the evacuation onset and decisive movement of occupants towards exits. Occupant reactions were on average slower when they were under the low-urgency treatment, when exit capacities were more restrictive and when the individual was located relatively far from the exit locations. Exit-choice observations showed that people’s behaviour fitted a multi-attribute trade-off that reflected the joint role of peer influence, choosing nearest exits and choosing visible exits. The shape of this trade-off remained, to large degrees, similar regardless of the urgency level. Observations of exit-choice adaptation showed that evacuees were two times more likely to change their exit choices when the simulated urgency level was higher. Further analysis revealed the joint effect of the queue-size imbalance at exits, exit visibility and peer influence on decision adaptation. The modelling outputs strikingly showed that evacuees exhibited notable follow-the-neighbour tendencies in their decision changing, whereas, they displayed the opposite tendency when making exit choices. The macro-scale analyses showed that the higher-urgency treatment triggered a more instant and faster rush to exits. This resulted in the formation of denser crowds at bottlenecks, but never prolonged the evacuation time. Evacuation times remained invariably shorter in high-urgency scenarios.