Abstract
In modern higher education institutions, the issue of excessive administrative burden has become critical. This often leads to the "digital paradox" phenomenon, where the implementation of centralized information systems, instead of streamlining operations, merely increases the time spent on reporting. This study proposes a solution to this problem through the development and empirical verification of a personal automated system for tracking faculty professional activity. The methodological framework for this work is the Design Science Research (DSR) approach and the "Citizen Developer" concept, which empowers educators without specialized technical training to create their own adaptive digital tools. The proposed architecture integrates the Telegram messenger, the Make.com platform as a scenario orchestration environment, and the Google Gemini 2.5 Flash multimodal artificial intelligence model. The developed algorithm facilitates complex routing for heterogeneous input data, including text descriptions, voice messages, web links (e.g., to scientific publications or news), and images (e.g., certificates or event posters). The system performs semantic content analysis, extracts key dates, generates descriptions in Ukrainian, and automatically logs entries into Google Sheets with linked digital copies of the relevant documents. Experimental validation conducted at the Faculty of Arts of Luhansk Taras Shevchenko National University (Poltava) demonstrated a reduction in the time required to record a single achievement from 4 minutes 31 seconds to 14 seconds, representing a 95% time saving. The paper analyzes the technical aspects of configuring system prompts to achieve a classification accuracy exceeding 90% and addresses information security concerns. The results demonstrate that the strategic use of accessible No-Code solutions enables the transformation of routine bureaucratic procedures into efficient processes, thereby freeing up faculty resources for scientific and creative endeavors.

This publication has 9 references indexed in Scilit: