Digitalization of management activities is accompanied by an increase in the number of social engineering threats implemented through corporate communication channels. Emotional attacks are particularly dangerous in this environment. These attacks are aimed at reducing the criticality of information perception, creating false trust, provoking urgency, guilt, fear, or expectation of benefit, and, as a result, encouraging the recipient to take actions contrary to the interests of the organization. Despite the active development of research in social engineering, cognitive decision-making mechanisms and intelligent management support systems, the problem of formalizing signs of emotional impact and their application in management decision support systems remains insufficiently developed. The article suggests a typology of emotional attacks, including trust, charm, community, good-naturedness, the prospect of benefits, lack of time, accusation and criticism. Four groups of formalized indicators of their recognition are identified. They are: linguistic, emotional, behavioral, and contextual ones. On this basis, a conceptual model of an intelligent system has been developed. The model provides the collection and preprocessing of digital communication data, the extraction of signs of emotional impact, risk assessment, interpretation of results and the development of recommendations for the head of the company or responsible specialist.