Pemanfaatan Gemini AI untuk Kreasi Konten Visual Promosi UMKM Kota Semarang

Abstract

This community service activity stems from a core challenge faced by Micro, Small, and Medium Enterprises (MSMEs) in Semarang City: the difficulty in consistently producing high-quality promotional visual content, primarily due to budget constraints and the lack of professional design skills. This condition severely hampers their competitiveness in the digital marketplace. The main goal of this activity was to enhance the technical capability of MSME participants in using Gemini AI as a multimodal model to generate professional, ready-to-use visual assets, thereby achieving significant cost and time production efficiency. The solution offered was a hands-on training focusing on Prompt Engineering defined as the art of formulating detailed commands to control the AI's output. The activity was conducted through an intensive workshop and case study mentorship for 25 MSME actors. The evaluation results demonstrated a highly significant increase in competence. The participants' average cognitive score drastically rose from 48.5% on the pre-test to 87.9% on the post-test, proving the successful mastery of Prompt Engineering. Applied results showed that participants were able to produce product visual assets with an average quality of 87.5%. Participant satisfaction reached 95.0%, especially concerning the indicators of cost and time efficiency. This activity successfully democratized design, empowering MSMEs to be self-sufficient in content production, and tangibly strengthening their brand identity and competitiveness across digital platforms.