Optimizing Marketing Materials Versioning Through Artificial Intelligence Solutions : improving resource efficiency at Kaleva Media
Hampinen, Lea Mae (2026)
2026
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202602183096
https://urn.fi/URN:NBN:fi:amk-202602183096
Tiivistelmä
Marketing on social media necessitates the continuous adaptation of content for various platforms, formats, and brand guidelines, making the versioning of campaign materials a time-consuming and repetitive process. This thesis focuses on the development and evaluation of an AI-assisted application aimed to support social media content adaptation and versioning at Kaleva Media.
The study followed an iterative, user-centered development approach. Platform-specific requirements for LinkedIn, Facebook, and Instagram were identified and combined with Kaleva Media’s regional brand guidelines. The developed application, built with React and python-based backend with FastAPI, integrates text generation with Google Gemini 2.5 Flash, image processing with Pillow (PIL), and video generation with OpenCV to produce branded static and animated social media content from user-provided text and images.
The application was evaluated through a structured user evaluation completed by a marketing professional at Kaleva Media. The evaluation assessed technical output quality, brand consistency, text generation quality, workflow efficiency, potential risks, and integration feasibility in relation to the research objectives. The results indicate that the application successfully meets platform and brand requirements, produces usable and coherent content variations, and significantly improves workflow efficiency compared to the manual process.
The findings suggest that AI-assisted tools can effectively support social media content production in media organizations by reducing repetitive work while maintaining brand consistency. The study also highlights the importance of continuous system monitoring and adaptation to evolving platform requirements. Future research could extend the evaluation to multiple users and examine long-term integration in everyday workflows.
The study followed an iterative, user-centered development approach. Platform-specific requirements for LinkedIn, Facebook, and Instagram were identified and combined with Kaleva Media’s regional brand guidelines. The developed application, built with React and python-based backend with FastAPI, integrates text generation with Google Gemini 2.5 Flash, image processing with Pillow (PIL), and video generation with OpenCV to produce branded static and animated social media content from user-provided text and images.
The application was evaluated through a structured user evaluation completed by a marketing professional at Kaleva Media. The evaluation assessed technical output quality, brand consistency, text generation quality, workflow efficiency, potential risks, and integration feasibility in relation to the research objectives. The results indicate that the application successfully meets platform and brand requirements, produces usable and coherent content variations, and significantly improves workflow efficiency compared to the manual process.
The findings suggest that AI-assisted tools can effectively support social media content production in media organizations by reducing repetitive work while maintaining brand consistency. The study also highlights the importance of continuous system monitoring and adaptation to evolving platform requirements. Future research could extend the evaluation to multiple users and examine long-term integration in everyday workflows.