Abstract: A method includes determining, without human intervention, statistically significant variables in a cluster of a first data set from variables in the first data set or second data sets. The cluster is associated with a consumer type. The method also includes generating pseudosentences based on high/low values for each statistically significant variable. The method also includes sending a signal to cause the pseudosentences to be inserted into a no-code prompt for a large language model to predict a segment description for the consumer type. The first no-code prompt has a code-style to reduce a variance of the segment description. The segment description is a representation of relationships among each statistically significant variable based on the set of pseudosentences. The method also includes sending, from the first compute device and to a second compute device, data representing the segment description.