Animal groups often exhibit enhanced capabilities that are outside the realm of a single individual. Our research aims to better understand what variables, processes and mechanisms control these beneficial emergent behaviours. Comprehension of these has exciting implications in understanding the building blocks of biological swarms and in the implementation of computerised applications and robotic platforms. Previous research has shown how simple, locally controlled rules of interaction can lead to robust group behaviours. Inspired by that work and from biological observations, we propose a novel behaviour selection algorithm that allows groups to effectively explore while maintaining unity. Using multi-agent computer simulations, we show that if individuals maintain close proximity with approximately six neighbours, the whole group can be coherent and mobile at the same time regardless of the group size and speed. Staying together as a coherent unit is a challenging task especially in fast moving groups and our algorithm, which is based on a simple sensory feedback, shows how these two often opposing behaviours can be consolidated to improve efficiency.