The development of software solutions relies on the documentation and digitization of existing business processes. Forms used in business processes are modelled by transferring existing form attributes to relational databases. Any change to the solution due to changes in business processes must ensure the consistency and security of existing data and requires a certain period to implement new business processes. Relational models are traditionally modelled abstraction of the real world to obtain the equivalent of a problem that can be solved programmatically. However, when changing business processes, it is not easy to change them due to the amount of data that relations contain. Data migration to new structures is required, with that data loss is possible due to insufficient planning or simply carelessness of the programmer. The introduction of new relational models, as well as the initial modelling of the process, requires the transfer of end users’ domain knowledge to the development engineer, and there can always be errors in communication during analysis, planning and execution of implementation. Dynamic models are one of possible solutions for this. Dynamic models transfer process modelling to the end user, and the software solution takes care of the data consistency related to the definition of the model (meta model). No migrations are needed, but simply a new definition of the model that will describe the new or improved business process. Following the definition of the model, the process of creating services for meta modelling and data entry using the Helm solution for managing a Kubernetes cluster, which are horizontally scalable, is presented. The subject of research in this work is the comparison of relational and dynamic models and the transfer of model creation to the end user, which would avoid upgrading the system and shorten the time of implementing new processes.