The paper addresses the automation of cognitive modelling for decision-making support in revealing economic sectors needs in digital technologies. It requires to take into consideration many factors, some of which have non-formalizable character. Some aspects do not have retrospective statistic history and are latent. Cognitive modelling is one of the artificial intelligence (AI) methods for describing such situations. It considers factors of enriching AI models by cognitive semantics, which experts help to create. The actual practice has shown that the process of model building can be costly and long term. An analysis of relevant big data (BD) can help to automate the process of cognitive modelling. There are two directions of automation: a) verification of cognitive models and b) synthesizing ones. The author’s convergent approach based on the inverse problem-solving method and the genetic algorithm for decision-making on the cognitive model was applied. It helps to make the process of cognitive modelling more purposeful and stable. An experimental test and actual practice application of the approach have shown a high level of the models’ verifications accuracy (about 93%) but a low level of accuracy of models synthesizing (about 34%). One of the reasons for such a low result lies in the fact that representatives of economic sectors and the digital industry often use different terminologies.