Komparace AI-asistovaného a AI-řízeného přístupu při vývoji business intelligence řešení

At present, artificial intelligence is influencing an increasingly wide range of industries. It has a significant impact especially on the field of information technology, where it is changing the way work is carried out through the rapid development of tools that support programming. This trend also applies to business intelligence and data engineering, where the demands placed on data specialists, the volume of processed data, and the complexity of the tasks being addressed are increasing. In this context, artificial intelligence shows the potential to partially reduce these demands. Nevertheless, there is only a limited number of studies focused on accelerating the development of business intelligence solutions through agentic development. This master’s thesis therefore provides insight into how artificial intelligence affects business intelligence, including the areas of data engineering and reporting. The thesis also includes the development of a business intelligence solution for a manufacturing company using two different approaches, which are subsequently compared within the solution development cycle. The first approach is traditional development supported by an AI assistant, while the second is AI-powered development using agentic artificial intelligence. The thesis shows that agentic development is already being implemented in a number of companies and brings results mainly in the areas of debugging and optimizing data pipelines, data integration, and data transformations. These findings also correspond with the results of the practical part of the thesis, where the greatest benefit was identified in the areas of data understanding, data engineering, the semantic layer, and reporting. AI-powered development also showed improvement in identifying data quality issues, particularly in the area of validity. Nevertheless, the role of the data specialist remains essential in the development process. However, their activities may gradually shift from technical tasks, such as setting up environments and writing transformation logic, toward working with business requirements, checking outputs, and adapting them to the needs of business users. Finally, a data architecture was proposed for the company that supports a higher-quality semantic layer for unifying metrics, the introduction of AI agents, and the potential use of conversational analytics.