Aspects of the disclosure relate to dynamically filtering and analyzing internal communications in an enterprise computing environment. A computing platform may receive communications data associated with electronic messages sent by various enterprise users. Subsequently, the computing platform may identify a subset of the communications data as being sent by a first enterprise user. Next, the computing platform may apply a grounded theory model to the subset of the communications data to produce a set of psycholinguistic parameter sub-scores, which the computing platform may map onto a plurality of predefined character traits to produce a user-specific character trait set. The computing platform may set a user-specific risk identifier based on the user-specific character trait set and may process communications data associated with other users. Then, the computing platform may apply an anomaly density evaluation and may produce and/or send a dataset that includes a prioritized list of high risk enterprise users.