The current (increasing) complexity of the economic activities, the significant volume of the data taken into consideration today at the level of a specific organization, the active competition environment have underlined the necessity of performance accounting and audit tools useful to answer to the “hunger” of information relevant to the users. Such a responsibility can not be accomplished by the classical audit tools, by themselves, fact that leads to the increased usage of the software tools specialized in audit.
In the Romanian environment, important applications in this area are not recognized and from this point of view, from this necessity, the research recommends a case study of multidimensional analysis of the accounting and financial data, integrating Data Mining techniques using Microsoft SQL Server 2005. These techniques can support auditor’s judgement offering complex and automatic analysis over a big set of data, discovering models, patterns or unnoticed tendencies. In this way, the research proposed the analysis of a set of data with a view to detect frauds, using Benford’s law, defining trends, patterns of the transactions using DMX language (Data Mining Extensions), analysing and identifying exceptions, calculating checking samples and the selection of those samples.