K-means is used to find clusters in a data space, where clusters group data points that share similar features with one another.