Capital budgeting models for project selection have been extensively explored, but relatively little attention has been given to the aspects of post-audit and project control for the projects implemented. This paper examines the area of project control and proposes a method that incorporates post-audit information as an active element in the decision to maintain or terminate an initiated project. Project selection is primarily based on anticipated performance, as determined from the most accurate information available. However, in a sequential decision environment, information that is accurate at one point in time may become inaccurate at another. This paper will show how the Dirichlet distribution can be utilized to formulate appropriately weighted prior probability beliefs, and how these initial beliefs can be updated as we receive post-audit information. To do this, we incorporate categorized cash flow data in a unique Bayesian-based framework. To illustrate the use of the Dirichlet distribution, we present a case study of an actual automation decision for a flexible manufacturing system. This case study also demonstrates how decision strategies can be improved by using post-audit information, when compared with conventional methods.