The volatility of the external economic environment actualizes the transition from deterministic economic and financial models to models with dynamic and stochastic properties. The purpose of the study is to improve the quality of management decisions in the field of company value management through the development and application of a graph stochastic model that ensures the adaptation of corporate governance to conditions of high volatility and uncertainty. Objectives of the study are as follows: to conduct a bibliographic analysis of the theory and practice of using Bayesian graph networks in the Russian economy; to develop an enlarged graph dynamic model of the formation of the value of a firm as a target indicator of corporate governance; on the basis of empirical data of a public company in the consumer sector to determine the stable structure and relationships of financial and economic indicators that form the value of the company; to test an enlarged graph stochastic model on empirical data; to calculate the aggregate distribution of the company’s value taking into account a set of influencing factors; to calculate the risks of positive and negative scenarios with a confidence level of the interval. An analysis of consumer sector company data for seven years was conducted, an enlarged graph dynamic model of value formation was developed, the parameters of the distribution of the values of the factors – nodes of the model were set. The data were simulated using the Monte Carlo method, the target value indicator was calculated in a probabilistic representation, the risks of positive and negative outcomes were assessed, scatter plots of influencing factors and the result were presented. To calculate and visualize the modeling results, a platform for stochastic modeling and interpretable AI based on Bayesian networks was used. Scientific novelty of the study is the use of Bayesian graph networks for cost engineering and upside risk assessment for the purposes of adaptive corporate governance, modeling of value in a probabilistic representation and with the calculation of an aggregate distribution based on empirical data of a consumer sector company, approbation of the approach of causal stochastic modeling to assess the probabilistic nature of the impact of shocks with a positive outcome on the cost, obtaining more realistic results in comparison with the static deterministic approach. Prospects for further research are the integration of the stochastic modeling approach into the practice of target engineering, the decomposition of factors – nodes of graph models, the development of cost models with an assessment of the influence of idiosyncratic factors in various industries.