People who are starting a business or considering self-employment often find the Czech tax
system difficult to navigate. They may lack a clear understanding of the available taxation
options and statutory obligations, which makes it harder to choose the most suitable tax
regime. The aim of this thesis is to develop a web application that simplifies the estimation
of taxes and mandatory contributions for self-employed individuals.
The application uses a cascading form structure, guiding users through the process step
by step while dynamically displaying only the options relevant to their situation. The form
is supported by a chat interface, where a contextual AI advisor provides targeted assistance
based on the question currently shown. By combining a cascading form with AI-based
guidance, the application reduces the risk of user errors and makes the process faster and
less cognitively demanding than existing solutions.
Based on the information entered by the user, the application compares the available
options and recommends the most advantageous solution. In cases where the entered income is close to a threshold affecting taxation or contributions, the application uses an
LLM to generate specific textual recommendations that go beyond the functionality of a
standard calculator. The application also partially supports voice-based form completion.
It is implemented using the Next.js framework and the TypeScript programming language.